Transcript
EMC® DiskXtender® File System Manager for UNIX/Linux Release 3.5
Administrator’s Guide P/N 300-009-573 REV. A01
EMC Corporation Corporate Headquarters: Hopkinton, MA 01748-9103 1-508-435-1000 www.EMC.com
Copyright © 2004-2009 EMC Corporation. All rights reserved. Published October 2009 EMC believes the information in this publication is accurate as of its publication date. The information is subject to change without notice. THE INFORMATION IN THIS PUBLICATION IS PROVIDED “AS IS.” EMC CORPORATION MAKES NO REPRESENTATIONS OR WARRANTIES OF ANY KIND WITH RESPECT TO THE INFORMATION IN THIS PUBLICATION, AND SPECIFICALLY DISCLAIMS IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Use, copying, and distribution of any EMC software described in this publication requires an applicable software license. For the most up-to-date listing of EMC product names, see EMC Corporation Trademarks on EMC.com. All other trademarks used herein are the property of their respective owners.
2
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Contents
Tables...................................................................................................................................7 Preface................................................................................................................................9 Chapter 1
Introduction File system-based storage management ........................................ 14 Files always available to applications......................................14 Data management...................................................................... 14 Console-based administration ........................................................ 16 Console Client ............................................................................ 16 Flexible topology............................................................................... 17 Minimum topology ................................................................... 17 Complex topology ..................................................................... 17 Topology guidelines.................................................................. 18
Chapter 2
Console overview Console software............................................................................... 20 Console Server............................................................................ 20 Console Agent ............................................................................ 20 Console Client ............................................................................ 21 Management domain ....................................................................... 22 Single computer domain............................................................22 Multiple computer domain .......................................................23 Console Client ................................................................................... 24 How to start Console Client ......................................................24 Administration .................................................................................. 26 How to add a user ..................................................................... 26 How to remove a user................................................................28
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
3
How to change your password ................................................ 28 How to change a user's password............................................ 29 Monitoring ......................................................................................... 30 Clearing the events window ..................................................... 30 Domain events ........................................................................... 31 Domain activities ....................................................................... 32 Host events ................................................................................. 34 File system events...................................................................... 35 File system activities ................................................................. 37 File system state ......................................................................... 38
Chapter 3
File System Configuration Preliminary tasks .............................................................................. 44 Back-end module types ............................................................. 44 Centera module preliminary tasks .......................................... 45 FTP module preliminary tasks ................................................. 52 NFS module preliminary tasks................................................. 54 Configuring a file system ................................................................ 57 Overriding configuration for file system subsets .................. 57 How to configure a file system................................................. 57 Centera module ......................................................................... 64 FTP module ................................................................................ 67 NFS module................................................................................ 68 Viewing a file system configuration .............................................. 70 Editing a file system configuration ................................................ 71 Deleting a file system configuration .............................................. 72
Chapter 4
Extended Rules Creating an extended rule ............................................................... 74 Match string................................................................................ 74 How to create an extended rule ............................................... 80 Viewing extended rules ................................................................... 86 Editing an extended rule ................................................................. 87 How to edit an extended rule ................................................... 87 Deleting an extended rule ............................................................... 88 How to delete an extended rule ............................................... 88 Copying extended rules................................................................... 89 How to copy extended rules ..................................................... 89
Chapter 5
Scheduled Jobs Scheduled jobs overview ................................................................. 92
4
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Effect on existing file system settings ......................................93 Delete Old Files job .................................................................... 94 Prune job...................................................................................... 95 Creating a scheduled job .................................................................. 96 How to create a scheduled job ..................................................96 Viewing a scheduled job ............................................................98 Editing a schedule...................................................................... 99 Deleting a schedule.................................................................. 100 Time elements.................................................................................. 102 Create a new time element ......................................................102 Editing a time element .............................................................107 Deleting a time element ...........................................................107
Chapter 6
Advanced UTDM Mount Options EMC Centera data retention .......................................................... 110 Performance impact..................................................................110 Requirements............................................................................ 111 Enabling EMC Centera data retention ...................................111 Setting retention for a file system ...........................................113 Setting retention for a group of files ......................................114 Restoring incremental backups...............................................115 Read-only ......................................................................................... 117 Mounting read-only................................................................. 118 Direct-read........................................................................................ 120 Memory-mapping.................................................................... 120 Mounting direct-read ...............................................................121
Chapter 7
Backup and Recovery What to back up............................................................................... 124 FSM-related files outside of UTDM file systems..................124 Files and data in UTDM file systems .....................................125 Backup software .............................................................................. 126 FSM-aware backup software...................................................127 Other backup software.............................................................127 Snapshot software.................................................................... 127 Recovering a lost file system ......................................................... 128 Recovering an FTP module file system........................................ 130 How to recover an FTP module file system ..........................130 EMC NetWorker.............................................................................. 132 Requirements............................................................................ 132 Backup with NetWorker software..........................................132
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
5
Recovery with NetWorker software ...................................... 133 NetWorker 64-bit Linux version................................................... 136 Requirements ............................................................................ 136 Backup........................................................................................ 137 Recovery .................................................................................... 137 NetBackup software ....................................................................... 139 NetBackup requirements......................................................... 139 Backup with NetBackup software ......................................... 141 Recovery with NetBackup software ...................................... 142 FSM tools.......................................................................................... 143 Backup with FSM tools ............................................................ 143 Recovery with FSM tools......................................................... 144 Snapshots with SnapView............................................................. 147 Quiescing a UTDM file system............................................... 147 Resuming UTDM file system activity ................................... 149 Recovering a UTDM file system snapshot............................ 149 Import data from SM to FSM ........................................................ 151
Chapter 8
Maintenance Processes .......................................................................................... 156 Core processes.......................................................................... 156 Console processes.................................................................... 162 File administration.......................................................................... 164 Manual migration, purge, and retrieval................................ 164 Periodic back-end system cleanup ............................................... 167 EMC Centera ............................................................................ 168 Cleanup on FTP or NFS module back-end systems ............ 170
Appendix A
Command reference Command quick reference ............................................................ Command usage ............................................................................. Setting the DMAP_ROOT_PATH ......................................... Setting the FSM environment ................................................ Administrative commands..................................................... File management commands ................................................. Backup and recovery commands ..........................................
174 177 177 177 178 191 196
Index ............................................................................................................................... 207
6
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Tables
Title 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
Page
Tasks permitted for operators and administrators .................................... Columns on the Events for All Hosts window ........................................... Columns on the All FSM Activities window .............................................. Columns on the Events windows ................................................................. Columns on the Events windows ................................................................. Columns on the Activities window .............................................................. File System State window .............................................................................. Purge tab on the File System State window ................................................ Migrate tab on the File System State window............................................. Staging tab on the File System State window ............................................. Destroy tab on the File System State window............................................. Recover tab on the File System State window ............................................ Available back-end module types ................................................................ Choices for the Authentication Type field................................................... Determining path of back-end mount point when using Solaris zones.. Determining full path when using Solaris zones ....................................... Checksum setting choices .............................................................................. Retention Period settings ............................................................................... Descriptions of Collision Avoidance settings ............................................. Possible Keyword values in match string expressions .............................. Possible Operator values in match string expressions .............................. Possible Variable values in match string expressions................................ Possible Filesize-Units values in match string expressions ...................... Operator precedence in a match string ........................................................ Descriptions of Checksum settings............................................................... Descriptions of Retention Period settings.................................................... Scheduled job comparison ............................................................................. Components of a schedule ............................................................................. Effect of a scheduled job on file system settings......................................... Delete Old Files job differences.....................................................................
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
27 32 33 35 36 38 39 40 40 40 41 41 45 53 55 58 59 64 66 75 75 77 78 78 83 84 92 93 94 95
7
Tables
31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66
8
Components of a time element ................................................................... Recurrence options ....................................................................................... Mounting with EMC Centera data retention ............................................ Mounting read-only...................................................................................... Mounting direct-read ................................................................................... Supported backup software ........................................................................ FSM recovery commands............................................................................. Mounting a file system for utdm_recdmattrf ........................................... Administrative commands .......................................................................... File management commands....................................................................... Backup and restore commands ................................................................... Options for deldmsession ............................................................................ Options for dx_read_log .............................................................................. Severity levels ................................................................................................ Brevity levels.................................................................................................. Options for dxuldmclip................................................................................ Options for dxuldmdelclips ........................................................................ Options for expand_dmattrf........................................................................ Options for getdmattr................................................................................... Options for getfileattr ................................................................................... Options for dxstat ......................................................................................... Options for dxprstat ..................................................................................... Options for dxhardlink................................................................................. Options for getmiglist................................................................................... Options for getmiglist................................................................................... getmiglist file list columns ........................................................................... Options for getpurgelist ............................................................................... getpurgelist file list columns headings ...................................................... Options for dxbuildtar ................................................................................. Options for dxcliplink .................................................................................. Options for dxdmload .................................................................................. Options for dxuldm_tar ............................................................................... Options for dmattrrecoverfs........................................................................ Options for getrecoverlist ............................................................................ getrecoverlist file list columns..................................................................... Options for utdm_fs_freeze .........................................................................
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
102 103 112 118 122 126 144 150 174 175 175 179 180 181 182 183 185 186 188 188 189 189 190 191 192 193 194 195 197 198 199 200 201 202 202 203
Preface
As part of an effort to improve and enhance the performance and capabilities of its product lines, EMC periodically releases revisions of its hardware and software. Therefore, some functions described in this document may not be supported by all versions of the software or hardware currently in use. For the most up-to-date information on product features, refer to your product release notes. If a product does not function properly or does not function as described in this document, please contact your EMC representative. Audience
This guide is part of the EMC DiskXtender File System Manager for UNIX/Linux, release 3.5 documentation set, and is intended for use by system administrators. Readers of this guide are expected to be familiar with the following topics: ◆
Their organization’s data archiving strategy, in particular: • Average archived file size. • Anticipated total number of files and total bytes to be archived. • File access norms: frequency of access during creation, edit, and archive phases. • Special file requirements, such as for shorter access times, multiple fail-safe copies, and optimal file stub size. • Backup strategies for data in the file systems.
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
9
Preface
◆
The network details of all computer systems and hardware devices to be used, including: • IP addresses • TCP/IP routing information • Switch, router, firewall, and network-attached device configurations
◆
Related Documentation
The administration details of all involved storage targets.
Related documents include: ◆
Release notes
◆
Installation guides: • AIX • HP-UX • Linux • Solaris • Console Client for Microsoft Windows • EMC AutoStart Module for EMC DiskXtender File System Manager for UNIX/Linux
The release notes contain a complete list of all related EMC product documentation. Conventions Used in This Guide
EMC uses the following conventions for notes and cautions. Note: A note presents information that is important, but not hazard-related.
!
IMPORTANT An important notice contains information essential to operation of the software.
10
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Preface
Typographical conventions EMC uses the following type style conventions in this document: Normal
Used in running (nonprocedural) text for: • Names of interface elements (such as names of windows, dialog boxes, buttons, fields, and menus) • Names of resources, attributes, pools, Boolean expressions, buttons, DQL statements, keywords, clauses, environment variables, functions, utilities • URLs, pathnames, filenames, directory names, computer names, filenames, links, groups, service keys, file systems, notifications
Bold
Used in running (nonprocedural) text for: • Names of commands, daemons, options, programs, processes, services, applications, utilities, kernels, notifications, system calls, man pages Used in procedures for: • Names of interface elements (such as names of windows, dialog boxes, buttons, fields, and menus) • What user specifically selects, clicks, presses, or types
Italic
Used in all text (including procedures) for: • Full titles of publications referenced in text • Emphasis (for example a new term) • Variables
Courier
Used for: • System output, such as an error message or script • URLs, complete paths, filenames, prompts, and syntax when shown outside of running text
Courier bold
Used for: • Specific user input (such as commands)
Courier italic
Used in procedures for: • Variables on command line • User input variables
<>
Angle brackets enclose parameter or variable values supplied by the user
[]
Square brackets enclose optional values
|
Vertical bar indicates alternate selections - the bar means “or”
{}
Braces indicate content that you must specify (that is, x or y or z)
...
Ellipses indicate nonessential information omitted from the example
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
11
Preface
Where to get help
EMC support, product, and licensing information can be obtained as follows. Product information — For documentation, release notes, software updates, or for information about EMC products, licensing, and service, go to the EMC Powerlink website (registration required) at: http://Powerlink.EMC.com
Technical support — For technical support, go to EMC Customer Service on Powerlink. To open a service request through Powerlink, you must have a valid support agreement. Please contact your EMC sales representative for details about obtaining a valid support agreement or to answer any questions about your account. Your comments
Your suggestions will help us continue to improve the accuracy, organization, and overall quality of the user publications. Please send your opinion of this document to:
[email protected]
12
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
1 Introduction
This chapter covers the following topics: ◆ ◆ ◆
File system-based storage management ......................................... 14 Console-based administration ......................................................... 16 Flexible topology................................................................................ 17
Introduction
13
Introduction
File system-based storage management EMC® DiskXtender® File System Manager for UNIX/Linux (FSM), release 3.5 is a data management solution that uses DMAPI-enabled (UTDM) file systems to transparently archive data on the following back-end storage systems: ◆
EMC Centera® Content Addressed Storage System (EMC Centera)
◆
Locally accessible disk resources
FSM transparently manages multiple local UTDM file systems. When necessary, it migrates and purges data from the file systems based on user-defined rules. FSM frees up space by transferring the actual data associated with a file system object to a back-end system, while retaining the object’s metadata in the file system. This means that a migrated and purged file is fully available through the file system, but uses very little of the file system’s space.
Files always available to applications When an application starts a read or write operation on data that has been migrated and purged, FSM invisibly triggers a retrieval of that data. The data is returned to the file system and the operation is completed. The user is not required to initiate operations to import or export data. Data written to an FSM file system is fully available for all read and write operations without user intervention. Access to migrated and purged data is the same as access to locally stored data, except for the possibility of a slight delay during the retrieval of data that has been purged from the file system. Information that can be provided from the file’s metadata, or from data contained in the locally retained stub file, is available without triggering a retrieval from the back-end system. This enables many queries to be completed at local disk speed, even for large files which have been migrated and purged.
Data management
14
FSM creates a transparent management layer on top of the native file system. The management layer uses a Data Management Application Programming Interface (DMAPI) to do the following:
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Introduction
◆
Interact with the file system.
◆
Manage extended file attribute information.
◆
Handle requests for migration, purging, and retrieval operations.
FSM migrates data to one or more back-end systems and manages data protection, retention, and replication on those systems. To fulfill requests for purged data, FSM retrieves the data from the appropriate back-end system.
File system-based storage management
15
Introduction
Console-based administration FSM provides a console-style administrative tool to perform most administrative tasks. The tool, the FSM Console, permits the administration of multiple UTDM file systems, UTDM file system hosts, and management domains. A management domain is a logical grouping of UTDM file system hosts controlled by the same FSM Console Server.
Console Client
FSM Console Client can be run on any supported host to administer a management domain, as long as it has TCP/IP access to the following management domain components: ◆
Console Server
◆
All UTDM file system hosts
Console Client can be run simultaneously on multiple hosts. You can use Console Client to perform tasks in the following categories: ◆ ◆ ◆ ◆
16
Administration Configuration Extended Rules Monitoring
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Introduction
Flexible topology FSM can be set up with many different combinations of hosts, file systems, protocols, and back-end systems. An FSM host is a computer where:
Minimum topology
Complex topology
◆
Core processes are installed and running.
◆
Console Agent is installed and running.
◆
At least one UTDM file system is available.
A minimal FSM installation involves the following: ◆
FSM host with Console Server installed and running.
◆
One UTDM file system.
◆
One back-end system.
A complex management domain might involve all of the following: ◆
One FSM host with Console Server installed and running.
◆
Multiple FSM hosts, including all of the supported operating systems.
◆
Multiple UTDM file systems on each FSM host, including all supported file system types.
◆
Multiple back-end systems, including EMC Centeras, and locally accessible disk resources.
◆
Additional EMC Centeras that act as replication servers.
◆
Console Client administration from multiple hosts.
Flexible topology
17
Introduction
Topology guidelines
18
To successfully set up a management domain, the following guidelines must be met: ◆
Only one Console Server may run in the management domain.
◆
All FSM hosts must have Console Agent installed and running.
◆
All FSM hosts must have TCP/IP communication with the Console Server host system.
◆
All Console Client hosts must have TCP/IP communication with the Console Server host and all FSM hosts.
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
2 Console overview
This chapter covers the following topics: ◆ ◆ ◆ ◆ ◆
Console software................................................................................ Management domain......................................................................... Console Client..................................................................................... Administration ................................................................................... Monitoring ..........................................................................................
Console overview
20 22 24 26 30
19
Console overview
Console software The Console software provides the interface to work with FSM hosts and UTDM file systems. An administrator can perform the following tasks by using the Console software: ◆
Configure new UTDM file systems.
◆
Edit the configuration of existing UTDM file systems.
◆
Delete the configuration information for existing UTDM file systems.
◆
Copy rule sets between UTDM file systems.
◆
Import and export rule sets.
◆
Create and manage extended rules.
The Console software consists of the following applications:
Console Server
◆
Console Server
◆
Console Agent
◆
Console Client
Console Server provides authentication services to a management domain. All FSM hosts in a management domain are controlled by the same Console Server. Console Server must be installed on only one computer in a management domain. That computer must have TCP/IP access to all FSM hosts in its domain. Normally the computer on which it is installed also has Console Agent installed. The computer on which Console Server is installed may also have Console Client installed.
Console Agent
20
Console Agent interacts with the processes and UTDM file systems on an FSM host. It acts as an intermediary between an FSM host and Console Client. It must be installed on each FSM host being managed.
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Console overview
Console Client
Console Client may be installed on any computer with TCP/IP access to the management domain. It provides an administrative interface to the management domain. Console Client may be active on several computers simultaneously and should be installed on every computer that is used to administer file systems. Console Client may run on a computer with or without Console Server and Console Agent.
Console software
21
Console overview
Management domain A management domain can exist entirely on a single computer or it can include multiple computers. A management domain consists of the following essential components: ◆
One Console Server.
◆
At least one Console Client.
◆
At least one FSM host. Each FSM host must have the following: • One set of FSM core processes (core processes) • One FSM Console Agent (Console Agent) • At least one UTDM file system
A UTDM file system is one that is first formatted as a native file system and then initialized to create a DMAPI attributes file in its root directory. The installation guides provide more information about this.
Single computer domain The following tasks must be completed to set up a management domain on a single computer: 1. Ensure the computer meets the installation requirements. 2. Install the core processes. 3. Install the Console Server. 4. Install a Console Agent. 5. Install the Console Client. 6. Prepare at least one UTDM file system. 7. Configure each UTDM file system.
22
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Console overview
Multiple computer domain A management domain can consist of many computers. All of the computers in the management domain must have TCP/IP communication with the following: ◆
Computer that is running Console Server
◆
Each computer that is running Console Client.
A multiple computer management domain might consist of the following separate computers: ◆
Console Server running on a single FSM host computer.
◆
Console Agent running on multiple FSM host computers.
◆
Console Client running on several computers.
Management domain
23
Console overview
Console Client Console Client is an interface application for working with FSM hosts and UTDM file systems. An encrypted username/password combination (account) is used to authenticate Console Client users and determine permissions. A default administrative account is provided with Console Client. This default account has the username "system" and the password "system". To increase the security of the management domain, the password for this account should be changed right after the software is installed. “How to change a user's password” on page 29 provides information about how to do this. You can install Console Client on several supported operating systems. The release notes provide a complete list of the operating systems supported by Console Client.
How to start Console Client To start Console Client: ◆
On UNIX and Linux hosts: 1. Allow X window connections: xhost +
2. Set the DISPLAY environment variable. setenv DISPLAY local_host:0.0
where local_host is the hostname of the local system or its IP address. Whether you use the local system’s hostname or its IP address depends upon the configuration of the network. The IP address can be either an IPv4 or IPv6 format IP address to match the protocol used by your network. 3. Run the Console Client application: /opt/fsm/client/File_System_Manager_Console_Client &
where /opt/fsm is the full path to the FSM installation directory. The Login to FSM Console window appears. 4. In Username, type a valid username. 5. In Password, type the password. 24
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Console overview
6. In FSM Server, type the hostname or IP address for the management domain’s Console Server. You can use either an IPv4 or IPv6 format IP address to match the protocol used by your network. 7. Click OK. The main Console Client window appears. ◆
On Windows: 1. Click: Start > Programs > File System Manager Console Client > File System Manager Console Client The Login to FSM Console window appears. 2. In Username, type a valid username. 3. In Password, type the password. 4. In FSM Server, type the hostname or IP address for the management domain’s Console Server. You can use either an IPv4 or IPv6 format IP address to match the protocol used by your network. 5. Click OK. The main Console Client window appears.
Security time-out
Online help
Console Client is configured to close down its connection with Console Server after 20 minutes of inactivity. This security feature cannot be configured. The online help provides convenient descriptions of Console Client’s menus, shortcuts, windows, and wizards. To access the online help from the main Console Client window, on the Help menu, click Contents and Index (or press F1). To access the online help from a wizard, click the Help button (or press Alt+F1).
Console Client
25
Console overview
Administration You can use the Console Client Administration menu to manage all user accounts in a management domain. The specific management domain is determined by the name of the Console Server host entered during login. Use the Administration menu to manage user accounts through the following commands:
!
◆
Manage User Accounts
◆
Change User Password
CAUTION The Console Client application installs with a default account that has the username "system" and the password "system". The role of this default account is Administrator. This account cannot be deleted. To maintain the security of the management domain, change the default password for the "system" as soon as the software is installed. “How to change a user's password” on page 29 provides more information.
How to add a user
To add a user: 1. On the Administration menu, select Manage User Accounts. 2. In Username, type the new username. A valid username is five to eight characters with at least one alphabetic character. A new username cannot match the following: • Any other username in the management domain. • The password of the currently logged-in user. 3. In Password, type the new password. A valid password is five to eight characters with at least one alphabetic character. It cannot match the password of the current logged-in user. 4. In Confirm Password, retype the new password. The value in Confirm Password must match the value in Password.
26
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Console overview
5. In Role, select a role for the new user account. Choose either Operator or Administrator. Refer to Table 1 on page 27 to view the functions available for each role. Tasks permitted for operators and administrators
Table 1
Task
Operator
Administrator
View all activities
Yes
Yes
View all events
Yes
Yes
Monitor file system
Yes
Yes
View file system configuration
Yes
Yes
View file system events
Yes
Yes
Change own password
Yes
Yes
View Help
Yes
Yes
View About
Yes
Yes
Manage user accounts
No
Yes
Create new file system configuration
No
Yes
Edit file system configuration
No
Yes
Export rule set
No
Yes
Import rule set
No
Yes
Commit changes
No
Yes
Delete file system configuration
No
Yes
Copy extended rule
No
Yes
6. Click Add. 7. Click OK. Note: Use Clear on the Manage User Accounts window to clear the following fields: Username, Password, and Confirm Password.
Administration
27
Console overview
How to remove a user To remove a user account: 1. On the Administration menu, select Manage User Accounts. The Manage User Accounts window appears. 2. From the Username list, select the username of the account to be removed. Note: The "system" account cannot be removed. Delete is not available when this account is selected.
3. Click Delete. The Delete User prompt appears. 4. Click Yes. The account is removed. 5. Click OK.
How to change your password To change your password: 1. On the Administration menu, select Change User Password. The Change Password window appears. 2. In Password, type your new password. 3. In Confirm Password, type the new password again. 4. Click OK.
28
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Console overview
How to change a user's password An administrator may change a user's password: 1. On the Administration menu, select Manage User Accounts. The Manage User Accounts window appears. 2. From the Username list, select a username. 3. In Password, type the new password. 4. In Confirm Password, type the new password. 5. Click Update. 6. Click OK.
Administration
29
Console overview
Monitoring To view events, activities, and status in a management domain, use the monitoring windows. The following regions may be monitored: ◆
Domain Events
◆
Domain Activities
◆
Host Events
◆
File System Events
◆
File System Activities
◆
File System State
Note: Events and state information for file systems on an FSM host are cleared when that FSM host is restarted. Restarting an FSM host also clears that host’s events from the Events for All Hosts window.
Clearing the events window An administrator can clear all events, or select events to clear from any of the Events windows. All Events To clear all events, click Clear All Events. All events clear, including events that are not displayed under the currently selected event type. Select Events To clear a group of events: 1. In the Filter field, select an event type to display. 2. Select each event to be cleared. 3. Click Clear Checked Events.
30
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Console overview
Domain events
The events for all file systems in a management domain may be viewed. The displayed events do not include the following: ◆
Events that were cleared.
◆
Events that occurred before the host computer was last restarted.
How to view domain events To display the events for a management domain: 1. On the View menu, select All Events. The Events for All Hosts window appears. 2. (Optional) On the View menu, select Refresh to refresh the display. The Last Refreshed value changes and the Events for All Hosts window updates. 3. (Optional) In the Filter field, select a severity level to display a subset of the events. The following severity levels may be selected: • All • None • Low • Medium • High • Extreme The filtered events appear. Table 2 on page 32 provides a description of the columns on the Events for All Hosts window.
Monitoring
31
Console overview
Table 2
Domain activities
Columns on the Events for All Hosts window Column
Description
Host
Name of the host on which the event occurred.
File System
Name of the file system that generated the event.
Severity
Severity level of the event. Severity is one of the following: • None • Low • Medium • High • Extreme
Timestamp
Date and time that the event occurred.
Event Type
Type of event. Event type is one of the following: • Debug • Info • Notice • Warning • Error • Critical • Alert • Emergency
Message
Description of the event.
The activities for all file systems in a management domain may be viewed. To display all activities in a management domain: 1. On the View menu, select All Activities. The All FSM Activities window appears at the top of the workspace pane. 2. In the Filter field, select the type of activities to display. The activities are filtered to show only the type of activities selected.
32
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Console overview
The following activities types may be selected: • All • Purge • Migrate • Staging • Destroy • Recover The filtered domain activities appear. Table 3 on page 33 provides a description of the columns on the All FSM Activities window. Table 3
Columns on the All FSM Activities window Column
Description
Host
Name of the host for the activity.
File System
Name of the file system for the activity.
File System ID
File system ID for the activity.
Backend Level
Priority of the back-end system. Either Level 1 (primary) or Level 2 (secondary).
Backend Type
Back-end module type, which is one of the following: • Centera • FTP • NFS
Activity Type
Activity type, which is one of the following: • Purge • Migrate • Stage • Destroy • Recover
Filename
Name of the file that is the subject of the activity.
Bytes to Move
Target amount of data (in bytes) to be moved by the activity.
Bytes Moved
Current amount of data (in bytes) that was moved by the activity.
Progress
Graphical progress bar.
Monitoring
33
Console overview
Host events
The events for an individual host may be viewed. The displayed events do not include the following: ◆
Events that were cleared.
◆
Events that occurred before the host computer was last restarted.
How to view events for a host To display all events for a host: 1. On the tree pane, select the host. 2. On the View menu, select Events. The Events window appears. 3. (Optional) On the View menu, select Refresh to refresh the display. The Last Refreshed value changes and the Events window updates. 4. (Optional) In the Filter field, select a severity level to display a subset of the events. The events are filtered to show only the selected severity level. The following severity levels may be selected: • All • None • Low • Medium • High • Extreme The filtered events appear. Table 4 on page 35 provides a description of the columns on the Events window.
34
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Console overview
Table 4
File system events
Columns on the Events windows Column
Description
Host
Name of the host.
File System
Name of the file system that generated the event.
Severity
Severity level of the event. Severity is one of the following: • None • Low • Medium • High • Extreme
Timestamp
The date and time that the event occurred.
Event Type
Type of event. Event type is one of the following: • Debug • Info • Notice • Warning • Error • Critical • Alert • Emergency
Message
Description of the event.
The events for an individual file system may be viewed. The displayed events do not include the following: ◆
Events that were cleared.
◆
Events that occurred before the host computer was last restarted.
How to view events for a file system To display all events for a file system: 1. On the tree pane, select the file system. 2. On the View menu, select Events. The Events window appears at the top of the workspace pane.
Monitoring
35
Console overview
3. (Optional) On the View menu, select Refresh to refresh the display. The Last Refreshed value changes and the Events window updates. 4. (Optional) In the Filter field, select a severity level to display a subset of the events. The events are filtered to show only the severity level selected. The following severity levels may be selected: • All • None • Low • Medium • High • Extreme The filtered events appear. Table 5 on page 36 provides a description of the columns on the Events window. Table 5
36
Columns on the Events windows (page 1 of 2) Column
Description
Host
Name of the host.
File System
Name of the file system that generated the event.
Severity
Severity level of the event. Severity is one of the following: • None • Low • Medium • High • Extreme
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Console overview
Table 5
File system activities
Columns on the Events windows (page 2 of 2) Column
Description
Timestamp
Date and time that the event occurred.
Event Type
Type of event. Event type is one of the following: • Debug • Info • Notice • Warning • Error • Critical • Alert • Emergency
Message
Description of the event.
You can use Console Client to view the activities for a file system. To display all activities for a file system: 1. On the tree pane, select the file system. 2. On the View menu, select Monitoring. The Activities window appears at the bottom of the workspace pane. 3. In the Filter field, select the type of activities to display. The activities are filtered to show only the type of activities selected. The following activities types may be selected: • All • Purge • Migrate • Staging • Destroy • Recover The filtered activities appear. Table 6 on page 38 provides a description of the columns on the Activities window.
Monitoring
37
Console overview
Table 6
File system state
Columns on the Activities window Column
Description
Activity Type
Type of activity. Type is one of the following: • Purge • Migrate • Stage • Destroy • Recover
Backend Level
Priority of the back-end system. Either Level 1 (primary) or Level 2 (secondary).
Backend Type
Back-end module type, which is one of the following: • Centera • FTP • NFS
Filename
Name of the file that is the subject of the activity.
Bytes to Move
Target amount of data (in bytes) to be moved by the activity.
Bytes Moved
Current amount of data (in bytes) moved by the activity.
Progress
Graphical progress bar.
You can use Console Client to view information about the state of a managed file system. To display the state of a file system: 1. On the tree pane, select the file system. 2. On the View menu, select Monitoring. The File System State window appears at the top of the workspace pane. General state information appears on the left side of the window. Table 7 on page 39 provides a description of the available information.
38
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Console overview
Table 7
File System State window Field Name
Description
File System Name Name of the file system. File System ID
File system ID.
Module
Module type for the file system.
File System Path
Full path to the mount point for the file system.
Total Size
Total size of the file system.
Used Space
Number of megabytes of data in the file system.
High Watermark
Value set for the file system's high-water mark.
Used Space (graphical display)
Graphical display showing the percentage of file system space used.
Maximum Inodes
Total number of inodes allowed in the file system.
Used Inodes
Total number of file system inodes used.
Used Inodes (graphical display)
Graphical display showing the percentage of file system inodes used.
3. (Optional) From the View menu, select Refresh to refresh the display. The Last Refreshed value changes, and the File System State window and the Activities window update. 4. On the File System State window, select a tab to display information for a specific activity type. The following tabs are available: • Purge, described in Table 8 on page 40. • Migrate, described in Table 9 on page 40. • Staging, described in Table 10 on page 40. • Destroy, described in Table 11 on page 41. • Recover, described in Table 12 on page 41.
Monitoring
39
Console overview
Table 8
Purge tab on the File System State window Field
Description
Purge Attempts
Number of attempts to purge files from the file system.
Purge Errors
Number of errors during all of the file system's purge attempts.
Purge Attempts on Purged Files
Number of purge attempts on previously purged files.
Bytes Purged Total bytes purged since the last time the file system's host was Since Last Reboot started. Table 9
Table 10
Migrate tab on the File System State window Field
Description
Migration Attempts
Number of attempts to migrate files from the file system.
Migration Errors
Number of errors during all of the file system's migration attempts.
Migration Attempts on Migrated Files
Number of migration attempts on previously migrated files.
Bytes Migrated Since Last Reboot
Total bytes migrated since the last time the file system's host was started.
Staging tab on the File System State window Field
Description
Staging Attempts Number of attempts to stage files into the file system. Staging Errors
Number of errors during all of the file system's staging attempts.
Staging Attempts Number of staging attempts on previously staged files. on Staged Files Bytes Staged Since Last Reboot
40
Total bytes staged since the last time the file system's host was started.
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Console overview
Table 11
Table 12
Destroy tab on the File System State window Field
Description
Files Destroyed
Total number of files reclassified on the back-end system as destroyed, after being deleted from the file system.
Files Not Renamed
Total number of files that could not be renamed on the back-end system. “Periodic back-end system cleanup” on page 167 provides more information about renaming deleted files on the back-end system.
Recover tab on the File System State window Field
Description
Recovery Attempts
Number of attempts to recover file data from the back-end system, after a recovery from backup is completed in the file system.
Recovery Errors
Number of errors during file data recovery attempts.
Monitoring
41
Console overview
42
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
3 File System Configuration
This chapter covers the following topics: ◆ ◆ ◆ ◆ ◆
Preliminary tasks................................................................................ Configuring a file system.................................................................. Viewing a file system configuration................................................ Editing a file system configuration.................................................. Deleting a file system configuration ...............................................
File System Configuration
44 57 70 71 72
43
File System Configuration
Preliminary tasks A UTDM file system’s configuration is a set of attributes that defines how it is managed. Each file system has its own set of required and optional attributes. Console Client simplifies file system configuration tasks by displaying all relevant attributes, checking all entered values, and applying configuration changes to the file system. Before you can configure a file system, you must complete the following: ◆
Create a native file system, initialize the file system for UTDM, and mount the file system on an FSM host. These tasks are described in the installation guides.
◆
Decide on a Level 1 back-end type and optionally, a Level 2 back-end type to use with the file system.
◆
Perform any preliminary setup steps required for each selected back-end system.
Back-end module types This release allows you to designate either one or two back-end systems for a file system. When two back-end systems are selected, migrated file data is written to both. Data is retrieved from the Level 1 (primary) back-end system unless it is unavailable. If the Level 1 back-end system is unavailable, then data is retrieved from the Level 2 (secondary) back-end system. The steps required to configure a file system differ slightly based on which back-end type is selected as primary and, optionally, which is selected as secondary. For both primary and secondary, an administrator may choose from the back-end module types shown in Table 13 on page 45.
44
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
File System Configuration
Table 13
Available back-end module types
Back-end module types
Description
Centera
EMC Centera
FTP
Standards-compliant FTP or FTPS servera
NFS
Standards-compliant NFS accessible disk resources
a. In order to facilitate recovery on an FTP or FTPS server, the server must be capable of providing FSM with local disk-like access to the back-end data. This may be provided using NFS or any other protocol that provides FSM with such access. “Recovering an FTP module file system” on page 130 describes the recovery procedure in this configuration.
A mixture of back-end module types can be used for a file system, for the file systems on a host, and for all file systems in a management domain. Each back-end module type has preliminary tasks that must be completed before you configure a file system to use it. Those tasks are described in the following sections: ◆
“Centera module preliminary tasks” on page 45
◆
“FTP module preliminary tasks” on page 52
◆
“NFS module preliminary tasks” on page 54
Centera module preliminary tasks The Centera module requires the following preliminary tasks: ◆
Ensure that the FSM host can establish a connection with the EMC Centera’s interface addresses. If replication is used, also confirm that there is a connection with the replication server’s interface addresses.
◆
Create a EMC Centera pool to use with FSM file systems.
◆
Configure the EMC Centera to grant the appropriate permissions to FSM. If replication is used, also configure the replication server.
◆
If Pool Entry Authorization (PEA) is used, create and install a valid PEA file on the FSM host. If replication is used, create a PEA file that authenticates FSM with both the primary EMC Centera and the replication server.
◆
Enable EMC Centera data retention. Preliminary tasks
45
File System Configuration
Check connectivity
FSM uses TCP/IP to communicate with EMC Centeras. Before configuring a UTDM file system to use a EMC Centera, and optionally a EMC Centera replication server, confirm that there is a connection between the FSM host and the IP addresses of the EMC Centeras. EMC Centeras do not respond to ping packets. A tool to confirm that there is a connection with a EMC Centera is provided with the FSM software. To confirm connectivity: 1. Log in as root on the FSM host system. 2. Set the FSM environment. To set the environment, refer to “Setting the FSM environment” on page 177. 3. For each EMC Centera IP address, run the dxuldmcenteraping command: dxuldmcenteraping -a emc-address [emc-address...]
where emc-address is an IP address for a network interface on the EMC Centera, in IPv4 format. Additional IP addresses, separated by spaces, may be specified. For more information about this command, see “dxuldmcenteraping” on page 182. Failure to connect to the IP addresses for a EMC Centera indicates a network problem that should be corrected before the file system is configured. Create a EMC Centera pool for FSM FSM file systems can use the default EMC Centera pool but it is recommended that you create a pool that is used only by FSM file systems. An FSM-only pool simplifies EMC Centera to EMC Centera migrations, such as when you move the back-end data to a newer generation EMC Centera. Configure EMC Centera permissions FSM requires read, write, and query permissions on a EMC Centera. This set of permissions can be provided in one of the following ways: ◆
46
Configure the EMC Centera’s Anonymous profile to grant these permissions.
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
File System Configuration
Use of the Anonymous profile eliminates the PEA file requirement. However, the Anonymous profile constitutes a potential security hazard since it does not authenticate applications during the PEA process. It should be disabled. ◆
Use the PEA process to create an application profile for FSM that grants the necessary permissions. For more information on this choice, see the next section: “Create and install a PEA file” on page 47.
Create and install a PEA file The PEA process is used by a EMC Centera to grant access rights to applications. The specific access rights that are granted are determined by application profiles created on the EMC Centera. The application profile that is used by FSM must have the following access rights: ◆
Read (r)
◆
Write (w)
◆
Exist (e)
In order to be authorized to use an application profile, FSM must provide the EMC Centera with the profile’s name and key. FSM reads the information from a PEA file located on the FSM host and passes it to the EMC Centera during the PEA process. To create and install a PEA file: 1. On the EMC Centera, create an application profile that grants Read, Write, and Exist rights. Application profiles are created by using the EMC Centera’s CLI commands profile create or profile update. In order to have the profile create or profile update commands output a PEA file, type yes when asked if a Pool Entry Authorization should be created. For more information about application profile creation, refer to the EMC Centera’s documentation. 2. Copy the resultant PEA file from the CLI host to the FSM host. 3. As root, limit access to the PEA file: chmod 600 /path/my.pea
where /path/my.pea is the full path of the PEA file. Preliminary tasks
47
File System Configuration
4. For each file system that uses the Centera module, type the full path of the PEA file in the Authentication Path option. Authentication Path is described in “Centera module” on page 64. Verify PEA access To verify that a PEA file can be used to access a EMC Centera: 1. Log in as root on the host system. 2. Set the FSM environment. For information on setting the FSM environment, see “Setting the FSM environment” on page 177. 3. Run dxuldmcenteraping: dxuldmcenteraping -a centeraIP?pathtopea
where: • centeraIP is one of the IP addresses on the EMC Centera, in IPv4 format. • pathtopea is the full path to the PEA file. Note: This command verifies that the PEA file may be used to access the EMC Centera. It does not check whether the correct permissions are provided by the PEA file.
For more information about dxuldmcenteraping, see “dxuldmcenteraping” on page 182. Centera pool entry authorization for replication The PEA process may be used with Centera replication as a way to authenticate FSM on both the primary and replication EMC Centeras. The PEA file used in a replication environment must provide authentication information for each of the EMC Centeras. The composition of a PEA file used in a replication environment depends on whether you provide the path to a random-bit file when you run the profile create or profile update commands. The random-bit file can be used as additional input during key generation. The following methods may be used: ◆
48
Generate the key by providing a random-bit file.
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
File System Configuration
The same file must be used on each EMC Centera. This method creates a PEA file with a single key that is accepted on each EMC Centera. ◆
Generate the key without providing a random-bit file. A different key is created on each EMC Centera based on various machine-based values. Each key must be copied into the PEA file. This creates a PEA file with several keys, one for each EMC Centera.
Replication PEA file created with a random-bit file
To use a replication PEA file created with a random-bit file: 1. On the primary EMC Centera, create an application profile by using a random-bit file. The application profile must grant Read, Write, and Exist rights. In order to have the profile create or profile update commands output a PEA file, type yes when asked if a Pool Entry Authorization should be created. For more information about application profile creation, refer to the EMC Centera’s documentation. 2. On each replication system, create an application profile by using the same random-bit file. The application profile on each replication system should use the same name as the one on the primary system. The profile should, at a minimum, grant Read and Exist rights. 3. Copy the PEA file created with the primary system from the CLI host to the FSM host. 4. As root, limit access to the PEA file: chmod 600 /path/my.pea
where /path/my.pea is the full path of the PEA file. 5. For each file system that uses the Centera module, type the full path of the PEA file in the Authentication Path option. This is described in “Configuring a file system” on page 57. 6. For each file system that uses the Centera module, type the IP address, in IPv4 format, for each available interface on each replication server in the Replication Server Addresses option. This is described in “Configuring a file system” on page 57.
Preliminary tasks
49
File System Configuration
7. Verify access using the PEA file. This task is described in “Verify PEA access” on page 48. Replication PEA file created without the use of a random-bit file
To use a replication PEA file created without the use of a random-bit file: 1. On the primary EMC Centera, create an application profile without using a random-bit file. The application profile must grant Read, Write, and Exist rights. In order to have the profile create or profile update commands output a PEA file, type yes when asked if a Pool Entry Authorization should be created. The EMC Centera’s documentation provides more information about application profile creation. 2. On each replication system, create an application profile without using a random-bit file. The application profile on each replication system should use the same name as the one on the primary system. The profile should, at a minimum, grant Read and Exist rights. 3. Copy the key portion of the PEA file created with each replication system into the PEA file created with the primary system, as shown in Example 1 on page 50. Example 1
Replication environment PEA file created without a random-bit file
The primary EMC Centera PEA file without the key from a replication server looks like this:
TXlQYXNzd29yZA== TXlQYXNzd29yZA==
The primary EMC Centera PEA file after adding the key from a replication server looks like this (added key is in bold font):
50
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
File System Configuration
TXlQYXNzd29yZA== TXlQYXNzd29yZA== YqgqywOqJ9nsKC7uQFeztUyFcM
4. Copy the resulting PEA file from the CLI host to the FSM host. 5. As root, limit access to the PEA file: chmod 600 /path/my.pea
where /path/my.pea is the full path of the PEA file. 6. For each file system that uses the Centera module, type the full path of the PEA file in the Authentication Path option. This is described in “Configuring a file system” on page 57. 7. For each file system that uses the Centera module, type the IP address, in IPv4 format, for each available interface on each replication server in the Replication Server Addresses option. This is described in “Configuring a file system” on page 57. 8. Verify access using the PEA file. This task is described in “Verify PEA access” on page 48. Enable EMC Centera data retention Retention of data from a UTDM file system that is migrated to a EMC Centera is enabled by using the Retention Period or the Retention Class settings. However, unless EMC Centera data retention is enabled for the file system, these retention settings do not protect data in the file system. To provide retention both on the back-end system and in the file system, mount the file system with the bckendreten mount option. For more information, see “EMC Centera data retention” on page 110.
Preliminary tasks
51
File System Configuration
FTP module preliminary tasks Use of the FTP module requires the following preliminary tasks: ◆
Ensure that the FSM host can establish a connection to the back-end server: • FTP server (FTP Protocol set to Native FTP) • FTPS server (FTP Protocol set to Native FTP)
Check connectivity
◆
Create a username/password combination on the back-end server for the file system.
◆
Obtain the full path to the target directory.
◆
Ensure that the file system host can obtain local disk-like access to the back-end server.
Connectivity between the FSM host and the back-end server can be confirmed by using ping. Use the ping command with the same interface address or hostname that you will use in FTP Host. To confirm connectivity: 1. Log in as root on the FSM host system. 2. Run ping: ping backendIP
where backendIP is the value to be used in FTP Host. This is normally the IP address of the back-end server, in either IPv4 or IPv6 as dictated by the network. Alternatively, if the back-end server’s hostname is to be used, replace backendIP with the hostname in the above command. Failure to connect to the IP address indicates a network problem that should be corrected before the file system is configured. Failure to connect to the hostname usually indicates a name service problem. FTP username and password FSM authenticates itself with the back-end server by using a username and password. This combination of values must be created on the back-end server before it can be used by a UTDM file system. Documentation for other FTP/FTPS servers provides information on creating FTP usernames and passwords.
52
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
File System Configuration
A single username/password combination can be used by all UTDM file systems, or individual combinations can be used by each FSM host or each UTDM file system. Target Path
Before configuring a file system, determine the full path of the directory on the back-end server that is to be the top-level migration target. Enter this path as the Target Path value for the file system. The Target Path value is described in “FTP module” on page 67. The path is in UNIX format and is relative to the root of the FTP server. Another commonly used target path naming convention is /FSM/file_system, where: ◆
/ is the root of the FTP directory structure on the FTP server.
◆
file_system is the name of the file system.
Choose an authentication type This release provides a choice between a standard FTP connection or a secure FTP connection for native FTP servers. IMPORTANT
!
The FTP module supports only active mode. When deciding on an authentication type be aware that using an encrypted control channel can prevent a firewall from opening a port for the data channel. Table 14 provides information about these choices. Choices for the Authentication Type field
Table 14
Authentication Type
Control channel encrypted
Data channel encrypted
Cryptographic protocol
None
No
No
None
TLS
Yes
No
Transport layer security
SSL
Yes
No
Secure socket layers
To successfully use either TLS or SSL the back-end server must provide a public key certificate.
Preliminary tasks
53
File System Configuration
Local disk-like access Local disk-like access is normally established by using the NFS protocol but other protocols are allowed, as long as the directory specified in the target path can be mounted locally by the FSM host.
NFS module preliminary tasks Use of the NFS module requires the following preliminary tasks:
Back-end mount point
◆
Determine the mount point of the back-end system.
◆
Create an NFS check directory on the back-end system.
The module can be used with a disk resource that acts as a back-end system, such as any of the following: ◆
Local secondary disk
◆
Network Attached Storage (NAS) disk
◆
Storage Area Network (SAN) disk
When the NFS module is used, determine the full path to the local mount point of the back-end system. Enter this path as the Target Path value for the file system. Configuring a file system’s Target Path value is described in “NFS module” on page 68. ◆
To mount a disk resource as a back-end system, refer to the instructions provided with the hardware.
Once the back-end system is mounted, the mount point can easily be determined by using mount. To determine the mount point: 1. Log in as root on the FSM host system. 2. On the command line, type mount without any options: mount
All mounted file systems are listed, with their mount point. Determining back-end mount point when using Solaris zones For Solaris zones the expression of the full path to the mount point of the back-end system depends upon the zone in which FSM is installed and the zone in which the back-end system is mounted. Table 15 on page 55 describes how to determine the full path when using Solaris zones.
54
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
File System Configuration
Table 15
Determining path of back-end mount point when using Solaris zones
Configuration
Path
• FSM installed in global zone • Back-end system is mounted in global zone
Use the global zone full path of the mount point.
• FSM installed in global zone • Back-end system is mounted in local zone
Use global representation of the local mount point. Example: Local zone is mounted at /export/zone01 and the back-end system’s local mount point is /mnt/backend01, then the full path to the mount point is /export/zone01/mnt/backend01.
• FSM installed in local zone • Back-end system is mounted in same local zone
Use local representation of the local mount point. Example: Local zone is mounted at /export/zone01 and the back-end system’s local mount point is /mnt/backend01, then the full path to the mount point is /mnt/backend01.
Create an NFS check directory The NFS check directory is an empty subdirectory at the top level of the back-end mount point. The NFS check directory must be named UTDMNFS, and be all uppercase. No specific permissions are required for the check directory. FSM uses the NFS check directory to ensure that the exported directory is properly mounted on the FSM Host. The NFS check directory can be seen by FSM only when the exported directory is properly mounted on the FSM host. FSM will not migrate files until it sees the NFS check directory. Note: Although a disk resource back-end system may not use the NFS protocol, the core processes will not migrate data unless the NFS check directory is found at the top level of the mount point.
To create an NFS check directory: 1. Log in as root on the FSM host system. 2. To ensure the back-end system is mounted, type mount without any options: mount
If the back-end system is not listed as one of the mounted file systems, then do the following: • For a disk resource back-end system, mount the resource as described in the instructions provided with the hardware. Preliminary tasks
55
File System Configuration
3. Change the current working directory to the top level of the back-end system. 4. Create the NFS check directory: mkdir UTDMNFS
56
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
File System Configuration
Configuring a file system Configuring a file system is the final step in the process of preparing to manage a UTDM file system. With this release you can configure a Level 1 and a Level 2 back-end system. When you configure both, migrated data is written to both. The Level 1 back-end system is the file system’s primary back-end system. Data is always retrieved from the Level 1 back-end system unless that back-end system is unavailable to the file system. If the Level 1 back-end system is unavailable, then data is retrieved from the Level 2 back-end system. When you configure two back-end systems for a file system you can select from any of the module types for each. Normally you should configure the back-end type that stages data fastest as the Level 1 back-end system.
Overriding configuration for file system subsets When you configure a file system you create settings that tell FSM how to manage the files in that file system. You can override those settings for subsets of the files using extended rules. “Extended Rules” on page 73 provides information about: ◆
Defining a file system subset
◆
Creating an extended rule
◆
Working with extended rules
How to configure a file system Configuring a file system: 1. Create, initialize, and mount a UTDM file system. The installation guides describe these procedures. 2. Start Console Client and log in to the management domain that controls the file system’s FSM host. “How to start Console Client” on page 24 describes how to start Console Client and how to log in to a management domain. 3. On the tree pane, select the file system’s host computer.
Configuring a file system
57
File System Configuration
4. Select File > New > File System Configuration. 5. Select the Level 1 (primary) back-end type from the following choices: Centera, FTP, or NFS. 6. (Optional) Check the Level 2 box. 7. (Optional) Select the Level 2 (secondary) back-end type. 8. Click OK. 9. In File System Name, type a reference name for the file system. The name may be any combination of ASCII characters, from 1 to 64 characters in length. The name cannot match the name of any other file system on the host. 10. In File System Path, type the path for the UTDM file system. Use the full path to the UTDM file system's mount point. For Solaris zones the expression of a file system’s full path depends upon the zone in which FSM is installed and the zone of the file system. Table 16 on page 58 describes how to determine the full path when using Solaris zones. Table 16
58
Determining full path when using Solaris zones
Configuration
Path
• FSM installed in global zone • File system installed in global zone
Use the global zone full path of the file system.
• FSM installed in global zone • File system installed in local zone
Use global representation of the local file system's path. Example: Local zone is mounted at /export/zone01 and the file system's local path is /mnt/filesystem01, then the path required is /export/zone01/mnt/filesystem01.
• FSM installed in local zone • File system installed in same local zone
Use local representation of the file system's path. Example: Local zone is mounted at /export/zone01 and the file system's local path is /mnt/filesystem01, then the path required is /mnt/filesystem01.
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
File System Configuration
11. In Checksum, select a checksum setting as described in Table 17 on page 59. Table 17
Checksum setting choices
Choice
Description
No checksumming
Checksums will not be generated. Use this setting to decrease the CPU requirements caused by data movement. This is the default setting because of its lower CPU requirements.
Checksum during migration
Checksum is generated for each file as it is migrated. This is the recommended setting. It creates a checksum for each migrated file without the additional CPU impact caused by creating a checksum during staging.
Checksum during migration and staging
Checksum is generated for each file as it is migrated. The checksum is regenerated and compared to the original as the file is staged. Use this setting if increased CPU impact is not an issue, if file integrity is critical, or to troubleshoot migration and staging problems. This setting is overridden by the staging setting of Partial Read Size. When that setting is enabled, a checksum is calculated only during migration.
When the value of Checksum is changed to either Checksum During Migration or Checksum During Migration and Staging, a checksum is not calculated for files that were migrated while the value of Checksum was No Checksumming. Checksums are calculated only for files that are migrated after the change. Note: Checksumming is disabled for back-end systems that use the Centera module and have Enable Multiple Streams Per File Stage selected.
12. Click Next. 13. (Optional) In Delay Until Purge Candidacy, set the number of minutes after a migrated file is last accessed before it can be purged. The range is 1 to 2,147,483,647 minutes. The default is 10 minutes. 14. (Optional) In Minimum File Size to Purge, set the minimum size a file must be before it can be purged. The range is 0 to 2,147,483,647 bytes (2 GB). The default is 0. 15. (Optional) In Sleep Interval Until Next Purge, set the number of minutes, after a system-initiated purge run completes, before the file system is again checked for files that can be purged.
Configuring a file system
59
File System Configuration
The range is 1 to 5,256,000 minutes. The default is 10 minutes. Note: When the sleep interval elapses, the file system is checked against its high-water mark. If the high-water mark is met or exceeded, a purge eligibility check is conducted.
16. (Optional) In Concurrent Files Purged, set the number of purge threads that can be started. The range is 1 to 128 files. The default is 1. Note: When the maximum number of threads is reached, files are queued until a thread becomes free.
17. (Optional) In File Stub Size, set the number of kilobytes of a file that are left in the file system when the file is purged. The range is 0 to 131,072 KB (128 MB). The default is 0. Note: The data is measured in kilobytes from the head of the file.
18. (Optional) In High Watermark, set the percent of file system capacity that is reached before automatic purging is started. The range is 1 to 100 percent, but must be higher than the value of low-water mark. The default is 90 percent. 19. (Optional) In Low Watermark, set the percent of file system capacity that is reached before automatic purging is stopped. The range is 1 to 100 percent, and must be less than the high-water mark. The default is 80 percent. 20. In the File Time Purge Policy section, select either Purge on Last Modification Time or Purge on Last Access Time. This setting determines the starting point for the Delay Until Purge Candidacy period: • Purge on Last Modification Time starts the period after the last modification of a file. • Purge on Last Access Time starts the period after the last time the file was accessed. 21. Click Next.
60
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
File System Configuration
22. (Optional) In Delay Until Migrate Candidacy, set the number of minutes after a file is last modified before it can be migrated. The range is 1 to 2,147,483,647 minutes. The default is 30 minutes. 23. (Optional) In Minimum File Size to Migrate, set the minimum number of bytes a file must be before it can be migrated. The range is 0 to 2,147,483,647 bytes (2 GB). The default is 0. 24. (Optional) In Sleep Interval Until Next Migration, set the number of minutes after a system-initiated migration run completes before the file system is again checked for files to migrate. The range is 1 to 5,256,000 minutes. The default is 30 minutes. 25. (Optional) In Concurrent Files Migrated, set the number of migration threads that can be started. The range is 1 to 128 files. The default is 8. Note: When the maximum number of threads is reached, files are queued until a thread becomes free.
26. (Optional) Set the Purge After Migrate: • Select to enable immediate purging of migrated files. • Clear it to enable purging only after the number of minutes set in Delay Until Purge Candidacy. 27. (Optional) Select Enable Data Compression and set a data compression level. When you enable this setting file system data is compressed before it is migrated. The compression level range is 0 to 9. Level 1 gives the fastest throughput, level 9 gives the greatest compression, and level 0 gives no compression at all (the input data is simply copied a block at a time). Level 6 represents a good compromise between speed and compression. The higher the compression you set the more the file system host's CPU will be utilized.
Configuring a file system
61
File System Configuration
Note: Selecting Enable Data Compression prevents the use of the staging settings: Partial Read Size and Full File Staging.
28. In the File Time Migration Policy section, select either Migrate on Last Modification Time or Migrate on Last Access Time. This setting determines the starting point for the Delay Until Migrate Candidacy period: • Migrate on Last Modification Time starts the period after the last modification of a file. • Migrate on Last Access Time starts the period after the last time the file was accessed. 29. Click Next. 30. (Optional) Clear Full File Staging, to set a partial read size. The default is full file staging, which means that all data for a file is returned to the file system when the file is staged. Full File Staging is disabled if Enable Data Compression is selected in the migration settings. 31. (Optional) In Partial Read Size, set the size in megabytes of file data chunks to return to the file system when a file is staged. Partial Read Size is disabled if Enable Data Compression is selected in the migration settings, or if Full File Staging is selected on the staging settings. The Checksum during Migration and Staging setting is overridden when Partial Read Size is enabled. A checksum is calculated only on files when they are migrated because partial staging prevents checksum calculations during staging. The range is 1 to 2048 MB (2 GB). The range of data staged spans the offset of the data requested. Data is retrieved in chunks from the highest "partial read size boundary" immediately before the requested data to the lowest "partial read size boundary" immediately after the requested data. Example 2 on page 63 provides further information about this.
62
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
File System Configuration
Example 2
Partial read size
A file is 1,073,741,824 bytes (1 GB). The file system’s Partial Read Size is set to 2,097,152 bytes (2 MB). The data requested is 4 MB in size and is located at offset 104,857,601 bytes (1 byte past the 100 MB block boundary). The data staged is the data from offset 104,857,600 bytes (100 MB is the highest "partial read size boundary" immediately before the requested data) to 111,149,056 bytes (106 MB is the lowest "partial read size boundary" immediately after the requested data). Note: Partial Read Size does not apply when the FTP module is used with the Native FTP protocol.
32. (Optional) In Concurrent Files Staged, set the number of staging threads that can be started. The range is 1 to 128 files. The default is 32. Note: When the maximum number of threads is reached, files are queued until a thread becomes free.
33. Click Next. 34. (Optional) In Concurrent Files Destroyed, set the number of threads that can be started to flag back-end data as deleted, for files that have been deleted from the file system. Deletion flags are described in “Periodic back-end system cleanup” on page 167. The range is 1 to 32 files. The default is 4. Note: When the maximum number of threads is reached, files are queued until a thread becomes free.
35. Click Next. 36. Complete the module-specific configuration for the selected Level 1 back-end system.
Configuring a file system
63
File System Configuration
Procedures for module-specific configurations appear in the following sections: • “Centera module” on page 64 • “FTP module” on page 67 • “NFS module” on page 68
Centera module
To complete the Centera module portion of a configuration: 1. In Connect Addresses, type the IP address or hostname of each available network interface on the EMC Centera. Use the standard IPv4 "dotted quad" format for each IP address (nnn.nnn.nnn.nnn) and separate each address by using a comma without a space. 2. (Optional) In Retention Period, select Do Not Set, None, Default, Infinite, or Days as described in Table 18 on page 64. If valid retention classes exist on the EMC Centera, Retention Period should be left at its default value, and Retention Class should be used to specify an appropriate retention class for the file system.
Table 18
Retention Period settings Setting
Description
Do not set Deleted data is retained for the period specified by the EMC Centera. This is the default and has the same result as selecting Default. None
Deleted data is immediately available for purging from the EMC Centera.
Default
Deleted data is retained for the period specified by the EMC Centera.
Infinite
Deleted data is never available for purging from the EMC Centera.
Days
Deleted data is not available for purging from the EMC Centera until after the specified number of days. If you select Days, then set the number of days. The range is 1 to 36,500 days.
3. (Optional) In Retention Class, type the name of a valid EMC Centera retention class. If valid retention classes exist on the EMC Centera, you can use Retention Class to enable retention of the file system’s data on the EMC Centera. This setting overrides the value of Retention Period. 64
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
File System Configuration
4. (Optional) In Replication Server Addresses, type the IP address of each of the replication server's available network interfaces. Use the standard, IPv4 "dotted quad" format for each IP address (nnn.nnn.nnn.nnn) and separate each address by using a comma without a space. 5. (Optional) In Centera Profile Path, type the full local path to the PEA file. “Create and install a PEA file” on page 47 provides more information about PEA files. 6. (Optional) In Buffer Size, clear Use Default and type a buffer size. The range is 1 to 10,240 KB (10 MB). The default value is 1 KB. Note: Buffer Size is the amount of memory allocated by the EMC Centera for temporary storage of data that requires further processing.
7. (Optional) In Timeout, clear Use Default and type the number of seconds before a file system operation times out. The range is 10 to 600 seconds (10 minutes). The default value is 10 seconds. Note: Timeout is the number of seconds the EMC Centera waits for an operation to finish before a timeout failure occurs.
8. (Optional) In Retry Count, clear Use Default and type a limit on the number of retries. The range is 0 to 99 retries. The default is the retry count set by the EMC Centera administrator. Note: Retry Count is the number of times to retry unsuccessful EMC Centera operations.
9. (Optional) In Retry Sleep, clear Use Default and type the number of seconds between retries. The range is 0 and 100 seconds. Note: Retry Sleep is the number of seconds the EMC Centera waits before retrying an unsuccessful operation.
Configuring a file system
65
File System Configuration
10. (Optional) In Embedded Data Size, clear Use Default and type the size of data that may be embedded in a C-Clip. The range is 0 to 65536 bytes (64 KB). The default is the size set by the EMC Centera administrator. Note: Embedded Data Size is the limit on the size of data BLOBs that are embedded in the EMC Centera C-Clip. Data BLOBs over the limit are referenced in the C-Clip, but stored elsewhere.
11. (Optional) In Prefetch Size, clear Use Default and type the size of memory on the EMC Centera to allocate for data transfers. The range is 1 to 1024 KB (1 MB). 12. (Optional) In Collision Avoidance, set how the EMC Centera should handle identical data BLOBs, as described in Table 19 on page 66. Table 19
Descriptions of Collision Avoidance settings Setting Description Default
Identical data BLOBs are handled according to the EMC Centera default setting.
On
Identical data BLOBs are handled separately, with unique ID numbers.
Off
Identical data BLOBs are handled as a single BLOB, with one ID number.
13. (Optional) Select Enable Multiple Streams Per File Stage, and set Streams Per File and Minimum Stream File Size. When Enable Multiple Streams Per File Stage is selected it allows multiple threads to be started for each staging request. This can enhance staging performance from an EMC Centera. Checksumming is fully disabled when this setting is selected. Set the number of threads in Streams Per File. The range is 1 to 32, however in this release no significant improvement has been shown for settings above 8. Set the minimum file size for which multiple threads are used in Minimum Stream File Size. Files that are less than this size will be staged using only one thread. The minimum size for this setting is 100 MB. The default value is 1024 MB.
66
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
File System Configuration
Note: Enable Multiple Streams Per File Stage is disabled for EMC Centera back-end systems that have Partial Read Size enabled, and when the file system has been mounted with the readthru option (direct-read capability).
14. Click Next. 15. (Optional) If you selected a Level 2 back-end system, complete the module-specific configuration related to it. Procedures for additional module-specific configurations appear in the following sections: • “FTP module” on page 67 • “NFS module” on page 68 16. On the Finish and Commit panel, click Finish. The configuration is validated, saved, and committed. The file system is ready to use and extended rules may be created.
FTP module
To complete the FTP module portion of a configuration: 1. In Target Path, type the full path of the target directory on the back-end system. “Target Path” on page 53 provides more information about the target directory. 2. In FTP Host, type the hostname or IP address of the back-end FTP server. The hostname must be 64 characters or less. The IP address can be either an IPv4 or IPv6 format IP address to match the protocol used by your network. 3. In FTP User ID, type the user ID to log in to the back-end FTP server. The user ID must be 64 characters or less. 4. In FTP Password, type the password to log in to the back-end FTP server. The password must be unencrypted, and must be 64 characters or less. 5. (Optional) In FTP Port, type the port that the back-end FTP server listens on. Configuring a file system
67
File System Configuration
The range is 1 to 65000. The default is port 21, the standard FTP and FTPS port. 6. (Optional) In Local Host IP Address, type a hostname, or IP address, for the file system host. This setting is available for hosts with multiple network interfaces. Use a hostname, or an IP address, for the selected network interface. The hostname must be 64 characters or less. The IP address can be either an IPv4 or IPv6 format IP address to match the protocol used by your network. The default uses the host's primary hostname and IP address. 7. In FTP Protocol, select Native FTP. Note: For Native FTP option, the following fields are unavailable: DXUL Family, Number of Instances, and Use Managed Disk Cache.
8. In Authentication Type, choose None, TLS, or SSL. “Choose an authentication type” on page 53 provides information about this choice. 9. Click Next. 10. (Optional) If you selected a Level 2 back-end system, complete the module-specific configuration related to it. Procedures for additional module-specific configurations appear in the following sections: • “Centera module” on page 64 • “NFS module” on page 68 11. On the Finish and Commit panel, click Finish. The configuration is validated, saved, and committed. The file system is ready to use and extended rules may be created.
NFS module
To complete the NFS module portion of a configuration: 1. In Target Path, type the full path for the local mount point of the target directory on the back-end system. “Back-end mount point” on page 54 provides more information.
68
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
File System Configuration
!
CAUTION The directory specified in Target Path must contain the NFS Check Directory or migration will not occur. “Create an NFS check directory” on page 55 provides more information about the NFS Check Directory. 2. Click Next. 3. (Optional) If you selected a Level 2 back-end system, complete the module-specific configuration related to it. Procedures for additional module-specific configurations appear in the following sections: • “Centera module” on page 64 • “FTP module” on page 67 4. On the Finish and Commit panel, click Finish. The configuration is validated, saved, and committed. The file system is ready to use and extended rules may be created.
Configuring a file system
69
File System Configuration
Viewing a file system configuration To view an existing file system’s configuration: 1. Start Console Client and log in to the management domain that controls the file system’s FSM host. The procedure to start Console Client and log in to a management domain is described in “How to start Console Client” on page 24. 2. On the tree pane, select the file system. 3. On the View menu, select Configuration. The View File System Configuration window appears at the top of the workspace pane. All attributes for the file system appear on the View File System Configuration window. The procedure “Configuring a file system” on page 57 describes each of the attributes.
70
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
File System Configuration
Editing a file system configuration To edit a file system configuration: 1. Start Console Client and log in to the management domain that controls the file system’s FSM host. The procedure to start Console Client and log in to a management domain is described in “How to start Console Client” on page 24. 2. On the tree pane, select the file system. 3. On the Edit menu, select File System Configuration. The File System Configuration wizard opens in edit mode. Attributes that cannot be edited are dimmed. 4. Make changes as needed and click Next to go to the next panel. The procedure “Configuring a file system” on page 57 describes each of the attributes. 5. On the Finish and Commit panel, click Finish. The edited configuration is validated, saved, and committed.
Editing a file system configuration
71
File System Configuration
Deleting a file system configuration To delete a file system configuration along with all of the extended rules assigned to the file system: 1. Start Console Client and log in to the management domain that controls the file system’s FSM host. The procedure to start Console Client and log in to a management domain is described in “How to start Console Client” on page 24. 2. On the tree pane, select the file system. 3. On the Edit menu, select Delete File System. The Delete File System prompt appears. 4. Click Yes. The file system configuration and all extended rules are deleted. Note: Deleting a file system configuration does not affect any of the following: the UTDM file system, data in the file system, or data migrated to the back-end system.
72
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
4 Extended Rules
This chapter covers the following topics: ◆ ◆ ◆ ◆ ◆
Creating an extended rule ................................................................ Viewing extended rules .................................................................... Editing an extended rule................................................................... Deleting an extended rule................................................................. Copying extended rules ....................................................................
Extended Rules
74 86 87 88 89
73
Extended Rules
Creating an extended rule Extended rules enable you to modify the data management policies for a select group of files in a file system. An extended rule overrides the file system’s configuration for the selected files. Affected files are determined by the match strings that you specify. The metadata for each file in the file system is compared to the match string to form the group of files affected by the extended rule. An extended rule is created by specifying the following:
Match string
◆
Criteria for selecting a set of files.
◆
Data management settings to apply to a set of files.
Every extended rule uses a match string to define the set of files covered by the rule. A match string is one or more expressions used to select a group of files. A match string can consist of several expressions separated by the AND operator and the OR operator. When the match string contains several operators, use parentheses to control the effects of operator precedence and associativity, as described in “Operator precedence and associativity” on page 78. A match string cannot exceed 127 characters, including space characters.
Parentheses
Parentheses are used in match strings to control the order in which operators are applied and expressions are evaluated. Parenthesis may be used with any of the following: ◆
One expression
◆
Several expressions
◆
Nested groups
One or more left-parenthesis can be added to the beginning of a match string, and after the AND and OR operators. You must add one or more right-parenthesis after an expression or expression group to close each open left-parentheses. Expressions
Match string expressions have the following format: KeywordOperatorVariable[Filesize-Units]
74
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Extended Rules
Keyword Keyword is described in Table 20 on page 75. Table 20
Possible Keyword values in match string expressions Keyword value Description FILENAME
Relative or full pathname To specify a particular directory use the full pathname including the mount point of the file system. After you enter this value it is placed in quotes by the software.
UID
User ID You can enter this value in quotes (string value) or without quotes (numeric value).
GID
Group ID You can enter this value in quotes (string value) or without quotes (numeric value).
FILESIZE
File size
Operator Operator is described in Table 21 on page 75. Table 21
Possible Operator values in match string expressions (page 1 of 2) Operator value
Description
==
Keyword is equal to Variable.
<>
Keyword is not equal to Variable.
>
Used only when the value of Keyword is FILESIZE, or when Keyword is UID or GID, and Variable is numeric. File size is greater than the Variable number of Filesize-Units, or UID or GID is greater than Variable.
Creating an extended rule
75
Extended Rules
Table 21
Possible Operator values in match string expressions (page 2 of 2) Operator value
Description
>=
Used only when the value of Keyword is FILESIZE, or when Keyword is UID or GID, and Variable is numeric. File size is greater than, or equal to, the Variable number of Filesize-Units; or UID or GID is greater than, or equal to, Variable.
<
Used only when the value of Keyword is FILESIZE, or when Keyword is UID or GID, and Variable is numeric. File size is less than the Variable number of Filesize-Units; or UID or GID is less than Variable.
<=
Used only when the value of Keyword is FILESIZE, or when Keyword is UID or GID, and Variable is numeric. File size is less than, or equal to, the Variable number of Filesize-Units; or UID or GID is less than, or equal to, Variable.
Variable The characteristics of Variable differ depending upon the context, as follows: ◆
String value when used with FILENAME (value you enter is placed in quotes by the software)
◆
String value when used with UID or GID and in quotes
◆
Integer value when used with UID or GID and not in quotes
◆
Integer when used with FILESIZE
The uses of Variable are described in Table 22 on page 77.
76
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Extended Rules
Table 22
Example 3
Possible Variable values in match string expressions Variable type
Description
String
When Keyword is FILENAME, UID, or GID. For UID or GID, a value is an integer unless it is in quotes. For FILENAME all values are strings and are placed in quotes by the software. Variable is a text string of 120 characters or less, representing a regular expression. Parentheses are not permitted. Wildcards may be used. However, the directory path symbol (/) must be included when creating a wildcard string expression that matches the beginning of the name of files that exist in more than one directory. This is described further in Example 3 on page 77.
Integer
When Keyword is FILESIZE, UID, or GID For UID or GID a value is an integer unless it is in quotes. When used with FILESIZE, it is an integer that represents the number of Filesize-Units. A FILESIZE comparison is formed by using the integer entered in Variable and the units selected in Filesize-Units. When not in quotes and used with UID or GID, it is an integer that represents a numeric user ID or group ID. The value of Variable must be numeric in this usage or the evaluation will fail.
Matching files in several directories by filename
You need an extended rule to apply to all files in a file system whose name starts with the letters "imp". The file system consists of many directories and subdirectories. In the extended rule wizard you select the FILENAME keyword and type the string "imp", without the quotes. The following appears: IF FILENAME=="imp*"
This string does not match the entire set. You select FILENAME and type the string */imp*. The following appears: IF FILENAME=="*/imp*"
This string matches the entire set.
Creating an extended rule
77
Extended Rules
Filesize-Units The Filesize-Units value is only used when the value of Keyword is FILESIZE. The possible values for Filesize-Units are described in Table 23 on page 78. Table 23
Possible Filesize-Units values in match string expressions Filesize-Units value
Description
bytes
Bytes
KB
Kilobytes
MB
Megabytes
GB
Gigabytes
Operator precedence and associativity When several operators exist in a match string, the operators are applied in the order of their precedence. If operators of equal precedence exist in the match string, they are applied by using the associativity rule. Placing an expression in parentheses causes that expression to be evaluated before the operator precedence and associativity rules are applied. Operator precedence and associativity are shown in Example 4 on page 79. Precedence Table 24 shows the precedence of match string operators. Table 24
Operator precedence in a match string Precedence order
Operator
First
< and <=
Second
> and >=
Third
== and <>
Fourth
AND
Fifth
OR
Associativity Operators of equal precedence are applied left-to-right. 78
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Extended Rules
Example 4
Operator precedence and associativity
Match String 1 (displayed): (FILENAME=="*data*" OR FILENAME=="*info*") AND UID<>123
The group of files defined by this match string contains both of the following: ◆
All files in the file system that do not have the user ID 123, and have the string "data" in their pathname. Note: This includes files whose pathname includes a directory with "data" in its name. The matching logic views a pathname as a single string expression. The character (/), normally consider a directory path character, is treated the same as any other character.
◆
All files in the file system that do not have the user ID 123, and have the string "info" in their pathname.
In Match String 1 the OR operator is applied before the AND operator because of the parentheses around the first two expressions. The expression "UID not equal to 123" ("UID<>123") is applied to the resulting two sets of files. Match String 2 (displayed): FILENAME=="*data*" OR FILENAME=="*info*" AND UID<>123
The group of files defined by this match string is different from the group defined by Match String 1. The Match String 2 group contains both of the following: ◆
All files in the file system that have the string "data" in their pathname.
◆
All files in the file system that do not have the user ID 123, and do have the string "info" in their pathname.
In Match String 2 the AND operator is applied before the OR operator. The expression "UID not equal to 123" ("UID<>123") is applied only to the set of files generated by the expression: FILENAME=="info"
Creating an extended rule
79
Extended Rules
How to create an extended rule To create an extended rule: 1. Start Console Client and log in to the management domain that controls the file system’s FSM host. “How to start Console Client” on page 24 describes how to start Console Client and how to log in to a management domain. 2. On the tree pane, select the file system that will use the extended rule. Note: Only a file system that is managed by FSM can use extended rules.
3. On the View menu, select Extended Rules. The file system's Extended Rules tab appears. 4. On the Extended Rules tab, click Create New Rule. The Extended Rule wizard appears. 5. In Rule Name, type a name for the new rule. The name may be any combination of ASCII characters, from 1 to 32 characters in length. The name cannot match the name of any other extended rule in the file system. 6. (Optional) In Migrate Candidacy Delay, specify a number of minutes after a matched file is last accessed before it can be migrated. The range is 0 to 2,147,483,647 minutes. The default is 30 minutes. 7. (Optional) In Purge Candidacy Delay, specify a number of minutes after a matched and migrated file is last accessed before it can be purged. The range is 0 to 2,147,483,647 minutes. The default is 10 minutes. 8. Form a match string: a. (Optional) Click the left-parenthesis to type one or more left-parenthesis. Parentheses are used to control operator precedence and associativity. b. Select one of the following keywords: FILENAME, UID, GID, FILESIZE. 80
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Extended Rules
The keyword defines the type of metadata used to define the matched set of files. Select one of the following operators: == , <> , > , >= , < , or <=. c. Type a value in the text box, as follows: – – – –
For the FILENAME keyword, type a string. For the UID keyword, type a quoted string, or an integer. For the GID keyword, type a quoted string, or an integer. For the FILESIZE keyword, type an integer.
Note: Wildcard characters may be used for keyword FILENAME, and for keywords UID and GID when a quoted string is used. The scale of the integer entered for keyword FILESIZE is determined by the size units.
d. (For keyword FILESIZE only) Select the size units: bytes, KB, MB, or GB. e. Click Add. The expression is added to the Match String. f. (Optional) Type one or more right-parenthesis to enclose an expression or expression group. The parentheses appear in the Match String. g. (Optional) Click AND or OR. The operator appears in the Match String. h. (Optional) Click NOT to modify an AND or OR operator. An exclamation mark appears following the AND or OR operator, changing them to AND NOT and OR NOT. i. (Optional) Repeat step a through step h to add another expression to the Match String. j. Click the right-parenthesis to close all open left-parentheses. The number of right-parentheses in the Match String must equal the number of left-parentheses.
Creating an extended rule
81
Extended Rules
Note: Click Undo to remove the latest element added to the Match String (an expression is considered a single element). Click Clear to remove all elements from the Match String.
9. Click Next. The Actions and Checksum panel appears. 10. (Optional) Select Do Not Migrate to prevent migration of matched files. 11. (Optional) Select Do Not Purge to prevent purging of matched files. 12. (Optional) Select Purge After Migrate to allow immediate purging of matched and migrated files. The default is to use the file system's Delay Until Purge Candidacy setting. 13. (Optional) Select Enable Data Compression and set a data compression level. When you enable this setting, file system data is compressed before it is migrated. Data compression level 1 is the least amount of compression and level 9 is the highest. The higher the compression you set the more the file system host's CPU will be utilized. Note: Selecting Enable Data Compression prevents the use of the staging setting Partial Read Size.
14. (Optional) In Stub Size, set the number of kilobytes retained in the file system for matched files. The range is 0 to 131,072 KB (128 MB). The default is 0. 15. In Checksum, select a checksum setting, as described in Table 25 on page 83.
82
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Extended Rules
Table 25
Descriptions of Checksum settings Setting
Description
Use file system checksum value
Checksums are generated as specified in the file system's Checksum setting. This is the default.
No checksumming
Checksums will not be generated for matched files.
Checksum during migration
A checksum is generated for each matched file when it is migrated.
Checksum during migration and staging A checksum is generated for each matched file when it is migrated. The checksum is regenerated and compared to the original when the file is staged.
16. In File Time Purge Policy, select either Purge on Last Modification Time or Purge on Last Access Time. This setting determines the starting point for the Delay Until Purge Candidacy period: • Purge on Last Modification Time starts the period after the last modification of a file. • Purge on Last Access Time starts the period after the last time the file was accessed. 17. In File Time Migration Policy, select either Migrate on Last Modification Time or Migrate on Last Access Time. This setting determines the starting point for the Delay Until Migrate Candidacy period: • Migrate on Last Modification Time starts the period after the last modification of a file. • Migrate on Last Access Time starts the period after the last time the file was accessed. 18. Click Next. 19. Complete the module-specific extended rule actions for the Level 1 back-end system.
Creating an extended rule
83
Extended Rules
Procedures for module-specific configurations appear in the following sections: • “Centera module actions” on page 84 • “NFS actions” on page 85 • “FTP actions” on page 85 20. (Optional) If you selected a Level 2 back-end system, complete the module-specific extended rule actions related to it. 21. Click Finish. The extended rule is validated, saved, and committed. Centera module actions To configure extended rule actions for a Centera module back-end system: 1. In Retention Period, select one of the settings described in Table 26. Table 26
Descriptions of Retention Period settings Setting
Description
Do Not Set Deleted data for matched files is retained for the period specified by the EMC Centera. This is the default and has the same result as selecting Default. None
Deleted data for matched files is immediately available for purging from the EMC Centera.
Default
Deleted data for matched files is retained for the period specified by the EMC Centera.
Infinite
Deleted data for matched files is never available for purging from the EMC Centera.
Days
Deleted data for matched files is not available for purging from the EMC Centera until after the specified number of days.
If you select Days, then set the number of days. The range is 1 to 36,500 days. 2. In Retention Class, type the name of a valid EMC Centera retention family to apply to matched files. 3. Click Next.
84
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Extended Rules
4. (Optional) If you selected a Level 2 back-end system, complete the module-specific extended rule actions related to it. 5. Click Finish. The extended rule is validated, saved, and committed. NFS actions
There are no extended rule actions specific to an NFS back-end system. All general actions are applied. To finish creating the extended rule: 1. Click Next. 2. (Optional) If you selected a Level 2 back-end system, complete the module-specific extended rule actions related to it. 3. Click Finish. The extended rule is validated, saved, and committed.
FTP actions
There are no extended rule actions specific to an FTP back-end system. All general actions are applied. To finish creating the extended rule: 1. Click Next. 2. (Optional) If you selected a Level 2 back-end system, complete the module-specific extended rule actions related to it. 3. Click Finish. The extended rule is validated, saved, and committed.
Creating an extended rule
85
Extended Rules
Viewing extended rules To view the extended rules for a file system: 1. Start Console Client and log in to the management domain that controls the file system’s FSM host. “How to start Console Client” on page 24 describes how to start Console Client and how to log in to a management domain. 2. On the tree pane, select the file system. 3. On the View menu, select Extended Rules. The file system's Extended Rules tab appears. The Extended Rules tab displays all rules for the file system. For each rule, the columns on the Extended Rules tab list the rule's general actions. The columns correspond to the actions described in “Creating an extended rule” on page 74. For each rule, the Extended Rules tab also displays any back-end system specific actions. The actions listed correspond to the actions described in: ◆
86
“Centera module actions” on page 84
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Extended Rules
Editing an extended rule An administrator may edit an extended rule.
How to edit an extended rule To edit an extended rule: 1. Start Console Client and log in to the management domain that controls the file system’s FSM host. “How to start Console Client” on page 24 describes how to start Console Client and how to log in to a management domain. 2. On the tree pane, select the file system. 3. On the View menu, select Extended Rule. The Extended Rules tab appears. 4. Select an extended rule to edit. 5. Click Edit Rule. The Extended Rule wizard opens in edit mode. Attributes that cannot be edited are dimmed. 6. Make changes as needed and click Next to go to the next panel. “Creating an extended rule” on page 74 provides more information about individual settings. Note: The Extended Rule wizard in edit mode is identical to the Extended Rule wizard, except that the match string Undo button is unavailable. The match string may not be edited by using Undo. To edit the match string, click Clear. This clears the existing match string so that a new match string can be created.
7. On the Finish and Commit panel, click Finish. The edited extended rule is validated, saved, and committed.
Editing an extended rule
87
Extended Rules
Deleting an extended rule An administrator may delete an extended rule. After deletion, the extended rule is no longer applied to the file system. Deleting an extended rule does not affect the following: ◆
The file system’s configuration information
◆
The file system’s data
How to delete an extended rule To delete an extended rule: 1. Start Console Client and log in to the management domain that controls the file system’s FSM host. “How to start Console Client” on page 24 describes how to start Console Client and how to log in to a management domain. 2. On the tree pane, select the file system. 3. On the View menu, select Extended Rule. The Extended Rules tab appears. 4. Select an extended rule to delete. 5. Click Delete Rule. A confirmation prompt appears. 6. Click Yes. The extended rule is deleted.
88
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Extended Rules
Copying extended rules An administrator may copy extended rules between file systems in the same management domain. This copies only the selected extended rules. File system configuration information and file system data are not copied.
How to copy extended rules To copy extended rules: 1. Start Console Client and log in to the management domain that controls the FSM hosts for the file systems. “How to start Console Client” on page 24 describes how to start Console Client and how to log in to a management domain. 2. From the Edit menu, select Copy Extended Rule. The Copy Extended Rule window appears. 3. In Copy From File System, select an FSM host and file system to copy from. 4. In Copy To File System, select an FSM Host and file system to copy to. Note: The file system selected in Copy To File System must use the same back-end module as the file system selected in Copy From File System.
5. In Extended Rule, select an extended rule to copy, or select Copy Entire Rule Set. Selecting Copy Entire Rule Set copies all of the extended rules for the selected file system. 6. Click OK to perform the copy operation. The selected extended rules are copied.
Copying extended rules
89
Extended Rules
90
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
5 Scheduled Jobs
This chapter describes schedules in the following topics: ◆ ◆ ◆
Scheduled jobs overview .................................................................. 92 Creating a scheduled job................................................................... 96 Time elements................................................................................... 102
Scheduled Jobs
91
Scheduled Jobs
Scheduled jobs overview Scheduled jobs allow you to define automatic run periods and blackout periods for the following file system jobs: ◆
Migrate
◆
Purge
◆
Delete Old Files
◆
Prune
For each managed file system, you can create one Migrate job and one Purge job. You can also create one Delete Old Files job and one Prune job for each of the file system’s back-end systems. Table 27 on page 92 compares each job before and after a scheduled job is created for it. Table 27
92
Scheduled job comparison
Job type
Without a scheduled job
With a scheduled job
Migrate
Runs automatically. Checks for migration candidates on the interval defined by the file system's Sleep Interval Until Next Migration setting.
Runs automatically during defined runtimes. Checks for migration candidates on the interval defined by the runtime element. Stops running during defined blackout times.
Purge
Runs automatically. Checks for purge candidates on the interval defined by the file system’s Sleep Interval Until Next Purge setting.
Runs automatically during defined runtimes. Checks for purge candidates on the interval defined by the runtime element. Stops running during defined blackout times.
Delete Old Files
Must be run manually as described in “Periodic back-end system cleanup” on page 167.
Runs automatically during defined runtimes. Deletes orphaned files on the back-end system (as described in “Delete Old Files job” on page 94) on the interval defined by the runtime element. Stops running during defined blackout times.
Prune
Will not run.
Runs automatically during defined runtimes. Only runs on the Level 1 back-end system, and only if the file system also has a Level 2 back-end system. Stops running during defined blackout times. Files on the Level 1 back-end system are “pruned” (marked as deleted) when the number of days since their migration exceeds the number of days set in the expiration period. Pruned files become available for removal by back-end system cleanup or Delete Old Files jobs. File system data is not affected by pruning and files are not removed from the Level 2 back-end system.
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Scheduled Jobs
A scheduled job consists of a standard file system job with an assigned schedule. The assigned schedule consists of the components described in Table 28 on page 93.
Table 28
Components of a schedule Component
Description
Name
Descriptive label for the schedule. The Name value must be unique for each schedule created for a file system.
Description
Information about the schedule's purpose.
Run Time
Time perioda during which the task is started regularly on the wake-up interval. More than one Run Time may be assigned to a schedule. Note: Blackout times have priority over runtimes. Jobs do not run during time periods in which blackout time segments coincide with runtime segments.
Black Out Time
Time period that the task cannot be run. More than one Black Out Time value may be assigned to a schedule.
a. Time periods are defined by the assigned time elements. Time elements are described in “Creating a scheduled job” on page 96.
Effect on existing file system settings A scheduled job can honor some file system settings and overrule others. Which settings are honored and which are overruled depends on the type of job scheduled, as described in Table 29 on page 94.
Scheduled jobs overview
93
Scheduled Jobs
Table 29
Effect of a scheduled job on file system settings
Job type
Honored settings
Overruled settings
Effect
Migrate
• Delay Until Migrate Candidacy • Minimum File Size to Migrate • Extended Rules
Sleep Interval Until Next Migration
Once a migrate job is scheduled, migration of eligible files is only started during the job's runtimes. No files are migrated during the job's blackout times. During runtimes the eligibility logic specified by the file system settings is applied. Extended rules are only applied during scheduled runtimes.
Purge
• Delay Until Purge Candidacy • Minimum File Size to Purge • Concurrent Files Purged • File Stub Size • High Watermark • Low Watermark • Concurrent Files Migrated • Extended Rules
• Sleep Interval Until Next Purge • Purge After Migrate
Once a purge job is scheduled, purge of eligible files is only started during the job's runtimes. No files are purged during the job's blackout times. A scheduled purge job can also overrule an enabled Purge After Migrate setting by preventing the immediate purging of files that are migrated during the purge job's blackout times. During runtimes the eligibility logic specified by the file system settings is applied. Extended rules are only applied during scheduled runtimes.
Delete Old Files
Concurrent Files Destroyed
None
None
Prune
None
None
None
Delete Old Files job
The Delete Old Files job removes files on a back-end system that have become orphaned because the files have been deleted from the file system. Those files are marked for deletion by FSM but are not automatically removed. Delete Old Files jobs have the same effect as the manual maintenance procedures described in “Periodic back-end system cleanup” on page 167. If a file system has two back-end systems, the Delete Old Files job only runs for the back-end system that is selected. The behavior of a scheduled Delete Old Files job changes depending on the type of back-end target. The differences are described in Table 30 on page 95.
94
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Scheduled Jobs
Table 30
Delete Old Files job differences Module type
Description
Centera
The Delete Old Files job applies to all eligible file data from all UTDM file systems that share the same virtual pool on an EMC Centera. Note: To avoid performance issues on the EMC Centera, only schedule a Delete Old Files job on one UTDM file system in any virtual pool. Since the job runs for all UTDM file systems in the virtual pool, scheduling more than one job wastes resources on the EMC Centera.
Prune job
FTP
The Delete Old Files job is not available for an FTP module back-end system. This task must be performed on the back-end host. “Periodic back-end system cleanup” on page 167 describes how to do this.
NFS
The Delete Old Files job applies to the selected file system only.
The Prune job removes files from a Level 1 back-end system after a set number of days (expiration period). This clears the Level 1 back-end system of older files, but the availability of all files is not affected. When files that have been pruned are needed they are retrieved from the Level 2 back-end system. Prune jobs have the following characteristics: ◆
Run only on Level 1 back-end systems
◆
Run only when the file system has both Level 1 and Level 2 back-end systems
◆
Cannot be run manually
◆
Files must be migrated before they are eligible for pruning
◆
Files do not need to be purged to be eligible for pruning
Scheduled jobs overview
95
Scheduled Jobs
Creating a scheduled job In overview, creating a scheduled job involves the following tasks: 1. Select a file system. 2. Select a job type: Migrate, Purge, Delete Old Files, or Prune. 3. If creating a Delete Old Files or Prune job, select a back-end system. 4. If creating a Delete Old Files and Prune job, set the Age or Expiration Period. 5. Create a named schedule. 6. Create time elements. 7. Add time elements to the Run Times and Black Out Times of the schedule. 8. Assign the schedule to the job. For each managed file system you can create the following jobs: ◆
One Migrate job
◆
One Purge job
◆
One Delete Old Files job for the Level 1 back-end system and one for the optional Level 2 back-end system
◆
One Prune job for the Level 1 back-end system only
Schedules that you create may be assigned to more than one scheduled job and may be used on more than one file system.
How to create a scheduled job To create a scheduled job: 1. On the tree pane, select the file system that will use the scheduled job. 2. On the View menu, select Schedules. The Schedules window is displayed with the file system listed in File System. 3. Select the Jobs tab, and in Job Type select a job type to assign to the schedule.
96
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Scheduled Jobs
4. (For Delete Old Files and Prune jobs only) In Backend, select the Level 1 or Level 2 back-end system. This setting only effects Delete Old Files and Prune jobs. Note: To schedule a Prune job the Level 1 back-end system must be selected. The Assigned Schedule section is not available, and a Prune job cannot be scheduled, if the Level 2 back-end system is selected (or if the file system only has a Level 1 back-end system).
5. (For Delete Old Files and Prune jobs only) Set a number of days: • Delete Old Files job—in Age (days), set the number of days that must pass, after the deletion of a file from the file system, before the file's data is eligible to be deleted from the selected back-end system. The default is 365 days. The range is 0 to 999,999 days. Note: Back-end system data can be used to restore files that are mistakenly deleted from the file system. To maximize this safeguard, set the Age value high to retain back-end data as long as possible.
• For a Prune job, in Expiration Period, set the number of days that must pass, after migration, before a file is pruned. The default is 365 days. The range is 1 to 999,999 days. 6. Select the Schedules tab, and in the Configured Schedules section click the Create a New Schedule icon. 7. On the Create a New Schedule box, enter a name for the schedule and click OK. Note: The schedule's name must be unique among all schedules on the file system's host. Use a descriptive name to make the schedule easier to recognize.
8. On the Schedules tab, in the Time Elements section, click the Add a New Time Element icon to open the Create a New Time Element wizard. 9. Use the Create a New Time Element wizard to create and save the time elements required for the new schedule. “Create a new time element” on page 102 describe how to use this wizard.
Creating a scheduled job
97
Scheduled Jobs
10. Select a time element to add as a runtime for the schedule, and click the Add the Selected Time Elements to the Current Run Times icon. Repeat this process to add any other time elements to Run Times. Note: Blackout times have priority over runtimes. Jobs do not run during time periods in which blackout time segments coincide with runtime segments.
11. Select a time element to add as a blackout time for the schedule, and click the Add the Selected Time Elements to the Current Black Out Times icon. Repeat this process to add any other time elements to Black Out Times. 12. Click Save. The new schedule is added to Configured Schedules. 13. In the Assigned Schedule section, click the Select a Schedule to Use icon. 14. On the Choose a Schedule box, select the name of the schedule you created, and click OK. The scheduled job is created and is assigned to the file system. It is graphically displayed in the monthly and daily calendars.
Viewing a scheduled job Scheduled jobs are graphically and textually displayed on the Jobs tab. To view the Jobs tab for a file system: 1. On the tree pane, select the file system. 2. On the View menu, select Schedules. The Schedules window is displayed. 3. Select the Jobs tab. The Jobs tab displays the schedules for each job type in the calendar month section.
98
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Scheduled Jobs
4. In Month and Year, select the month and year that you wish to view. The calendar month displays the selected month. 5. In the calendar month section, click on a day of the month. The job schedules for that day are graphically portrayed in the day view, and textually displayed in the Time Spans view.
Editing a schedule
Editing a schedule involves the following tasks: ◆
Remove all assigned tasks from the schedule.
◆
Remove time elements that need to be changed from the schedule's Run Times and Black Out Times.
◆
Edit time elements and create new time elements.
◆
Add time elements to the schedule's Run Times and Black Out Times.
◆
Save the schedule.
To edit a schedule: 1. On the tree pane, select the file system that uses the schedule. 2. On the View menu, select Schedules. The Schedules window appears. 3. On the Jobs tab, in the Selected Jobs section, select a Job Type that is assigned to the schedule. 4. In the Assigned Schedule section, click the Revert Back icon. Repeat steps 3 and 4 for each Job Type that is assigned to the schedule. 5. On the Schedules tab, in the Configured Schedules section, select the schedule. 6. In the schedule's Run Times section, select a time element to remove or edit and click Remove. Repeat this for each time element that you wish to remove or edit. 7. In the schedule's Black Out Times section, select a time element to remove or edit and click Remove. Repeat this step for each time element to remove or edit.
Creating a scheduled job
99
Scheduled Jobs
8. In the Time Elements section, click the Add a New Time Element icon to create a new time element, or select an existing time element and click Edit the Selected Time Element to edit it. Both actions open the Create a New Time Element wizard. Use the wizard to configure and save the time element as described in “Create a new time element” on page 102. Repeat this for each time element to add to the edited schedule. 9. Select a time element to add as a runtime for the schedule, and click the Add the Selected Time Element to the Current Run Times icon. Repeat this process to add any other time elements to Run Times. 10. Select a time element to add as a blackout time for the schedule, and click the Add the Selected Time Element to the Current Black Out Times icon. Repeat this process to add any other time elements to Black Out Times. 11. Click Save. The new schedule is added to Configured Schedules and is ready to use. To complete the schedule a task must be assigned. This is described in “How to create a scheduled job” on page 96.
Deleting a schedule
Deleting a schedule involves the following tasks: ◆
Removing all assigned tasks from the schedule
◆
Deleting the schedule
To delete a schedule: 1. On the tree pane, select the file system that uses the schedule. 2. On the View menu, select Schedules. The Schedules window appears. 3. On the Jobs tab, in the Selected Jobs section, select a job type that is assigned to the schedule.
100
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Scheduled Jobs
4. In the Assigned Schedule section, click the Revert Back icon. Repeat steps 3 and 4 for each job type that is assigned to the schedule. 5. On the Schedules tab, in the Configured Schedules section, select the schedule and click the Delete the Selected Schedule icon. The schedule is deleted.
Creating a scheduled job
101
Scheduled Jobs
Time elements Time elements are the building blocks of a schedule. Before you can create a scheduled job you must create at least one time element. Table 31 on page 102 describes the components of a time element. Table 31
Components of a time element Component
Description
Recurrence
The calendar cycle that provides the framework for the schedule. This can be daily, weekly, monthly, yearly, weekends, or weekdays.
Wake-up interval
Time interval on which the task is started. The task is started at each wake-up interval during the runtime. However, if the task is already running from a previous wake-up, additional processes are not started. Note: The wake-up interval is determined based on the host’s internal clock, not the last start or finish of the job. For example, a 15 minute wake-up interval with a daily recurrence would check if the job is running and start it if is not at: midnight, 12:15 A.M., 12:30 A.M., 12:45 A.M., and so forth.
Start time
Beginning of the time period included in a time element.
End time
End of the time period included in a time element.
“Creating a scheduled job” on page 96 describes how to create a scheduled job by using the available time elements.
Create a new time element To create a new time element: 1. On the Schedules tab, in the Time Elements section, click the Add a New Time Element icon. The Create a New Time Element wizard launches. 2. Select one of the recurrence options. Table 32 on page 103 describes the options.
102
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Scheduled Jobs
Table 32
Recurrence options Name
Description
One Specific Time Starts a job on a specific date and time and does not recur. Daily
Starts a job every day at a specific time.
Monthly
Starts a job on a specific day of the month at a specific time and recurs each month.
Weekends
Starts a job at a specific time on Saturday and Sunday and recurs every Saturday and Sunday.
Weekly
Starts a job on a specific day of the week at a specific time and recurs every week.
Yearly
Starts a job on a specific month, day of the month, and time and recurs every year.
Weekdays
Starts a job at a specific time on each weekday (Monday-Friday) and recurs every weekday.
3. Click Next. The fields on the wizard's second screen change depending upon the recurrence option selected. To complete the creation of a time element go to the relevant section: • “One Specific Time” on page 104 • “Daily” on page 104 • “Monthly” on page 105 • “Weekends” on page 105 • “Weekly” on page 105 • “Yearly” on page 106 • “Weekdays” on page 106
Time elements
103
Scheduled Jobs
One Specific Time
To complete a time element that uses the One Specific Time recurrence option: 1. In Wake Up Interval, specify the number of minutes that should pass between checks to see if the job is running and, if it is not, start the job. During specified runtime, the job is checked on the wake-up interval to see if it is running. If the check shows that the job has finished, it is started again. The wake-up interval is determined from the host’s internal clock, not from when the job last started or finished. 2. In Year, specify the year for the time element. 3. In From, specify the month, day, and time to start. This is the one time at which the time element starts the job running. 4. In Until, specify the month, day, and time that the time element stops running, or select Do not stop. A Do not stop selection configures the time element to start at the start time and run until the end of the year. 5. Click Finished.
Daily
To complete a time element that uses the Daily recurrence option: 1. In Wake Up Interval, specify the number of minutes that should pass between checks to see if the job is running and, if it is not, start the job. 2. In From, specify a start time. This is the time every day at which the time element starts the job running. 3. In Until, specify the time that the time element stops running, or select Do not stop. A Do not stop selection configures the time element to start at the start time and run until the end of the day. 4. Click Finished.
104
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Scheduled Jobs
Monthly
To complete a time element that uses the Monthly recurrence option: 1. In Wake Up Interval, specify the number of minutes that should pass between checks to see if the job is running and, if it is not, start the job. 2. In From, specify the day of the month and time to start. This is the time every month at which the time element starts the job running. 3. In Until, specify the day of the month and time that the time element stops running, or select Do not stop. A Do not stop selection configures the time element to start at the start time and run until the end of the month. 4. Click Finished.
Weekends
To complete a time element that uses the Weekends recurrence option: 1. In Wake Up Interval, specify the number of minutes that should pass between checks to see if the job is running and, if it is not, start the job. 2. In From, specify a start time. This is the time every Saturday and Sunday at which the time element starts the job running. 3. In Until, specify the time that the time element stops running, or select Do not stop. A Do not stop selection configures the time element to start at the start time and run until the end of the day. 4. Click Finished.
Weekly
To complete a time element that uses the Weekly recurrence option: 1. In Wake Up Interval, specify the number of minutes that should pass between checks to see if the job is running and, if it is not, start the job. 2. In From, specify the day of the week and time to start. This is the time every week at which the time element starts the job running.
Time elements
105
Scheduled Jobs
3. In Until, specify the day of the week and time that the time element stops running, or select Do not stop. A Do not stop selection configures the time element to start at the start time and run until the end of the week. 4. Click Finished. Yearly
To complete a time element that uses the Yearly recurrence option: 1. In Wake Up Interval, specify the number of minutes that should pass between checks to see if the job is running and, if it is not, start the job. 2. In From, specify the month, day, and time to start. This is the time every year at which the time element starts the job running. 3. In Until, specify the month, day, and time that the time element stops running, or select Do not stop. A Do not stop selection configures the time element to start at the start time and run until the end of the year. 4. Click Finished.
Weekdays
To complete a time element that uses the Weekdays recurrence option: 1. In Wake Up Interval, specify the number of minutes that should pass between checks to see if the job is running and, if it is not, start the job. 2. In From, specify a start time. This is the time every weekday at which the time element starts the job running. 3. In Until, specify the time that the time element stops running, or select Do not stop. A Do not stop selection configures the time element to start at the start time and run until the end of day. 4. Click Finished.
106
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Scheduled Jobs
Editing a time element To edit a time element: 1. On the Schedules tab, in the Time Elements section, select the time element. 2. Click the Add a New Time Element icon. The Create a New Time Element wizard launches with selected time element’s settings. 3. Use the wizard to edit the time element settings and click Finished.
Deleting a time element Deleting a time element involves the following tasks: ◆
Removing the time element from each schedule in a file system.
◆
Deleting the time element.
To delete a time element: 1. On the tree pane, select the file system that uses the time element. 2. On the View menu, select Schedules. The Schedules window is displayed. 3. On the Jobs tab, in the Selected Jobs section, select a Job Type that is assigned to a schedule that uses the time element. 4. In the Assigned Schedule section, click the Revert Back icon. 5. Repeat steps 3 and 4 for each job type that is assigned to a schedule that uses the time element. 6. Repeat steps 3, 4, and 5 for each schedule that uses the time element. 7. On the Schedules tab, in the Configured Schedules section, select the schedule. 8. If the time element is in the schedule's Run Times section, select it and click Remove. 9. If the time element is in the schedule's Black Out Times section, select it and click Remove.
Time elements
107
Scheduled Jobs
10. Repeat steps 7, 8, and 9 for each schedule that uses the time element. 11. On the Schedules tab, in the Time Elements section, select the time element and click the Delete the Selected Time Element icon.
108
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
6 Advanced UTDM Mount Options
This chapter describes the following mount options: ◆ ◆ ◆
EMC Centera data retention........................................................... 110 Read-only .......................................................................................... 117 Direct-read ........................................................................................ 120
Advanced UTDM Mount Options
109
Advanced UTDM Mount Options
EMC Centera data retention EMC Centera data retention uses the bckendreten mount option to provide a mechanism to prevent the following data operations in the file system and on the back-end system: ◆
Overwrite
◆
Truncation
◆
Removal
◆
Modification of its metadata
◆
Modification of its retention period
These operations cannot be initiated in the file system or on the EMC Centera by any process, including non-FSM processes, for files that have been migrated or that are eligible for migration. A file is eligible for migration if it has not been modified for at least the period specified in the file system’s Delay Until Migrate Candidacy value. When a file is subject to EMC Centera data retention, the file’s data cannot be changed during the retention period, in the file system or on the EMC Centera. Data on a secondary back-end system is not protected by this data retention mechanism.
Performance impact UTDM file systems that are mounted with the bckendreten mount option generate a retention event to the core processes when a file modification operation is attempted. This event determines whether the modification is allowed. The core processes instruct the UTDM file system to deny the attempt if either of the following is true for a file: ◆
The file is migrated and there is an unexpired retention period for its data on the EMC Centera.
◆
The file is not migrated and has not been modified within the period specified in the file system’s Delay Until Migrate Candidacy value. There is also a Centera Retention Period or Centera Retention Class associated with the file.
The retention events can have a performance impact on file modification operations. A slight impact can occur because the core
110
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Advanced UTDM Mount Options
processes must query the EMC Centera for the applicable retention period. Retention events will not adversely impact normal file creation operations. UTDM file systems do not generate a retention event during the first minute after a file is created or modified.
Requirements
EMC Centera data retention requires the following: ◆
The EMC Centera must use one of the following license models: • Basic Compliance • Compliance Plus
◆
The UTDM file system must use the EMC Centera as its primary back-end system.
◆
The UTDM file system configuration must specify at least one of the following: • A value in Retention Class • A value in Retention Period other than None • An extended rule that sets a retention period or retention class
◆
The file system must be mounted with the bckendreten mount option.
◆
The FSM host system and the Centera should be in the same timezone for retention to be in sync.
Enabling EMC Centera data retention To set EMC Centera data retention periods, first enable retention for the file system. Retention is specified by using the bckendreten option when mounting the file system. To enable EMC Centera data retention for a file system: 1. Log in to the file system’s host as root. 2. Set the FSM environment. “Setting the FSM environment” on page 177 for more information. 3. Stop all FSM processes: kill.fsm
EMC Centera data retention
111
Advanced UTDM Mount Options
4. If the file system is mounted, unmount it: umount [-f] /file-system
where /file-system is the full path to the file system’s mount point. Note: On Linux systems, the -f option must be used when you unmount a UTDM type file system. This is because the file system is viewed as busy by the operating system.
5. Mount the file system by using one of the methods shown in Table 33 on page 112. The variables for the procedures shown in Table 33 are as follows: • fstype is the native type of the file system. • blkdev is the full path to the block special device. • mntpnt is the full path to the mount point. • chardev is the full path to the character special device. • log is the log file that is generated for the logical volume. Table 33
112
Mounting with EMC Centera data retention (page 1 of 2)
Mounting method
Procedure
Command line (All operating systems, except Linux and AIX)
As root, run the following mount command:
Command line (Linux only)
As root, run the following mount command:
Command line (AIX only)
As root, run the following mount command:
mount-F utdm -o fstype=fstype,bckendreten /blkdev /mntpnt
mount-t utdm -o fstype=fstype,bckendreten /blkdev /mntpnt
mount-V utdm -o fstype=fstype,bckendreten /blkdev /mntpnt
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Advanced UTDM Mount Options
Table 33
Mounting with EMC Centera data retention (page 2 of 2)
Mounting method
Procedure
File system table (Solaris only)
As root, edit the /etc/vfstab file so that the entry for the file system uses the bckendreten option, as shown here: /blkdev /chardev /mntpnt utdm 2 yes fstype=fstype,bckendreten
Save the file and mount the file system: mount /mntpnt
File system table As root, edit the /etc/fstab file so that the entry for the file system uses the (Linux and HP-UX only) bckendreten option, as shown here: /blkdev /mntpnt utdm fstype=fstype,bckendreten 0 2
Save the file and mount the file system: mount /mntpnt
File system table (AIX only)
As root, edit the /etc/filesystems file so that the options line of the file system’s entry includes the bckendreten option: options=rw,fstype=fstype,log=/dev/log,bckendreten
Save the file and mount the file system: mount /mntpnt
6. Restart the FSM processes: rc.fsm
EMC Centera data retention is enabled for the file system.
Setting retention for a file system To set EMC Centera data retention for all files in a file system: 1. Mount the file system by using the bckendreten mount option. “Enabling EMC Centera data retention” on page 111 provides more information. 2. Start Console Client and log in to the management domain that controls the file system’s FSM host. “How to start Console Client” on page 24 describes how to start Console Client and log in to a management domain. 3. On the tree pane, select the file system. 4. On the Edit menu, select File System Configuration. The File System Configuration wizard opens in edit mode. 5. Click Next until the Level 1 - Centera Backend panel appears. EMC Centera data retention
113
Advanced UTDM Mount Options
6. In Retention Class, type the name of an EMC Centera retention class. Note: If a suitable retention class does not exist on the EMC Centera, you can specify a retention period. However, Retention Period cannot be None if Retention Class is empty.
7. Click Next until the Finish and Commit panel appears. 8. Click Finish. The configuration is validated, saved, and committed. The file system is ready to use and EMC Centera data retention is active. Note: Steps 1 through 8 describe how to edit an existing file system configuration to set EMC Centera data retention. You may edit other file system configuration parameters when performing these steps. You can also set EMC Centera data retention when a file system is first configured.
Setting retention for a group of files To set EMC Centera data retention for a group of files in a file system: 1. Mount the file system by using the bckendreten mount option. “Enabling EMC Centera data retention” on page 111 provides more information. 2. Start Console Client and log in to the management domain that controls the file system’s FSM host. “How to start Console Client” on page 24 describes how to start Console Client and log in to a management domain. 3. On the tree pane, select the file system. 4. On the View menu, select Extended Rules. 5. On the Extended Rules tab, click Create New Rule. The Extended Rule wizard appears. 6. In Rule Name, type a name for the new rule. The name may be any combination of ASCII characters, from 1 to 32 characters in length. The name cannot match the name of any other extended rule in the file system.
114
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Advanced UTDM Mount Options
7. Form a match string that identifies the group of files. “Creating an extended rule” on page 74 provides information about forming a match string. 8. Click Next until the Centera Actions panel for the primary back-end system appears. 9. In Retention Class, type the name of a valid EMC Centera retention class. Note: If a suitable retention class does not exist on the EMC Centera, you can specify a retention period. However, Retention Period cannot be None if Retention Class is empty.
10. Click Next until the Finish and Commit panel appears. 11. Click Finish. The extended rule is validated, saved, and committed. The file system is ready to use and EMC Centera data retention is active for the specified group of files. Note: Steps 1 through 11 describe how to create an extended rule to set EMC Centera data retention for a group of files. You may specify other extended rule parameters when performing these steps. You can also edit an existing extended rule to add EMC Centera data retention.
Restoring incremental backups EMC Centera data retention can prevent the use of incremental backups to restore some files to a UTDM file system. The affected files are those that have all of the following characteristics: ◆
The file is not a migration candidate when the full backup is created.
◆
The file is modified after the full backup is created and before incremental backup is created.
◆
The time between the creation of the full backup and the restoring of the incremental backup is greater than the value of the Delay Until Migrate Candidacy value.
When all of these characteristics are true for a file, the UTDM file system allows the restore of the file from the full backup, but rejects the restore of the file from the incremental backup. EMC Centera data retention
115
Advanced UTDM Mount Options
To restore incremental backups while using EMC Centera data retention: 1. Start Console Client and log in to the management domain that controls the file system’s FSM host. “How to start Console Client” on page 24 describes how to start Console Client and log in to a management domain. 2. On the tree pane, select the file system. 3. On the Edit menu, select File System Configuration. The File System Configuration wizard opens in edit mode. Attributes that cannot be edited are unavailable. 4. Click Next until the Optional Migration Settings panel appears. 5. Increase the value of Delay Until Migrate Candidacy. The value of Delay Until Migrate Candidacy is expressed in minutes. It must be increased to be greater than the number of minutes from the time of the full backup to the time of the restore. 6. Click Next until the Finish and Commit panel appears. 7. Click Finish. The edited configuration is validated, saved, and committed. 8. Perform a restore by using the incremental backup. 9. Change Delay Until Migrate Candidacy back to its original value. Restore of the incremental backup is complete.
116
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Advanced UTDM Mount Options
Read-only This release provides the utdmro mount option to permit a UTDM file system to be mounted read-only. A read-only mounted UTDM file system functions the same as a standard read-only mounted file system, with one exception: when the UTDM read-only mount option is used, the .DMATTR file is allowed to function normally. This exception is required to permit the file system to exchange data with the back-end system and perform DMAPI processing. A read-only, mounted UTDM file system permits the following file system activity: ◆
Reading files
◆
Reading files by using direct-read
◆
Directory traversals
◆
Migration Files that are ready to migrate after the file system is mounted read-only will not be migrated.
◆
Purge
◆
Staging
◆
Backup
◆
Recovery processing To perform a recovery, mount the file system normally with Read/Write permissions and perform the complete recovery of the file system before the file system is remounted in read-only mode.
!
IMPORTANT A UTDM file system must be mounted read-write before a recovery can be performed. A read-only mounted UTDM file system does not allow: ◆
File deletions
◆
File permissions or attribute changes
◆
File data modification or truncations Read-only
117
Advanced UTDM Mount Options
Mounting read-only
To mount a UTDM file system read-only: 1. Log in to the file system’s host as root. 2. Set the FSM environment. “Setting the FSM environment” on page 177 for more information. 3. Stop all FSM processes: kill.fsm
4. If the file system is mounted, unmount it: umount [-f] /file-system
where /file-system is the full path to the file system’s mount point. Note: On Linux systems, the -f option must be used when you unmount a UTDM type file system. This is because the file system is viewed as busy by the operating system.
5. Mount the file system by using one of the methods shown in Table 34 on page 118. The variables for the procedures shown in Table 34 on page 118 are as follows: • fstype is the native type of the file system. • blkdev is the full path to the block special device. • mntpnt is the full path to the mount point. • chardev is the full path to the character special device. • log is the log file that is generated for the logical volume. Table 34
118
Mounting read-only (page 1 of 2)
Mounting method
Procedure
Command line (All operating systems, except Linux and AIX)
As root, run the following mount command:
Command line (Linux only)
As root, run the following mount command:
Command line (AIX only)
As root, run the following mount command:
mount-F utdm -o fstype=fstype,utdmro /blkdev /mntpnt
mount-t utdm -o fstype=fstype,utdmro /blkdev /mntpnt
mount-V utdm -o fstype=fstype,utdmro /blkdev /mntpnt
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Advanced UTDM Mount Options
Table 34
Mounting read-only (page 2 of 2)
Mounting method
Procedure
File system table (Solaris only)
As root, edit the /etc/vfstab file so that the entry for the file system uses the bckendreten option, as shown here: /blkdev /chardev /mntpnt utdm 2 yes fstype=fstype,utdmro
Save the file and mount the file system: mount /mntpnt
File system table As root, edit the /etc/fstab file so that the entry for the file system uses the (Linux and HP-UX only) bckendreten option, as shown here: /blkdev /mntpnt utdm fstype=fstype,utdmro 0 2
Save the file and mount the file system: mount /mntpnt
File system table (AIX only)
As root, edit the /etc/filesystems file so that the options line of the file system’s entry includes the bckendreten option: options=rw,fstype=fstype,log=/dev/log,utdmro
Save the file and mount the file system: mount /mntpnt
Read-only
119
Advanced UTDM Mount Options
Direct-read This release provides the readthru mount option to enable direct-read on a UTDM file system. Direct-read enhances file system performance by delivering file data from the back-end system directly to a requesting application’s buffers without first writing the data into the file system. Direct-read works with multiple back-end systems. If the primary back-end system cannot be read, the direct-read request is made of the secondary back-end system. Note: Direct-read does not work if NFS is configured as Level 1 and Centera as the Level 2 back-end system.
The direct-read mount option can be used on any UTDM file system. However, only the following back-end modules support it: ◆
NFS
◆
Centera
When a file system that is mounted with direct-read requests data from an unsupported back-end system the data is staged into the file system and then transferred to the application’s buffers. Direct-read applies only to reading a file. If the file’s data is altered in any way, then the file is staged into the file system as it would be without direct-read. Note: The readthru mount option does not work for pruned and compressed files.
Memory-mapping
Applications that use the mmap function call to perform memory mapping of file data cannot benefit from direct-read. When a application uses this function the back-end data is staged into the file system as it would be without direct-read. In particular, Solaris uses the mmap function call in many of its system utilities, such as cat and cp. The GNU versions of these utilities do not use the mmap function and allow direct-read to work.
120
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Advanced UTDM Mount Options
Mounting direct-read To mount a UTDM file system direct-read: 1. Log in to the file system’s host as root. 2. Set the FSM environment. “Setting the FSM environment” on page 177 for more information. 3. Stop all FSM processes: kill.fsm
4. If the file system is mounted, unmount it: umount [-f] /file-system
where /file-system is the full path to the file system’s mount point. Note: On Linux systems, the -f option must be used when you unmount a UTDM type file system. This is because the file system is viewed as busy by the operating system.
5. Mount the file system by using one of the methods shown in Table 35 on page 122. The variables for the procedures shown in Table 35 are as follows: • fstype is the native type of the file system. • blkdev is the full path to the block special device. • mntpnt is the full path to the mount point. • chardev is the full path to the character special device.
Direct-read
121
Advanced UTDM Mount Options
• log is the log file that is generated for the logical volume. Table 35
Mounting direct-read
Mounting method
Procedure
Command line (All operating systems, except Linux and AIX)
As root, run the following mount command:
Command line (Linux only)
As root, run the following mount command:
Command line (AIX only)
As root, run the following mount command:
File system table (Solaris only)
As root, edit the /etc/vfstab file so that the entry for the file system uses the readthru option, as shown here:
mount-F utdm -o fstype=fstype,readthru /blkdev /mntpnt
mount-t utdm -o fstype=fstype,readthru /blkdev /mntpnt
mount-V utdm -o fstype=fstype,readthru /blkdev /mntpnt
/blkdev /chardev /mntpnt utdm 2 yes fstype=fstype,readthru
Save the file and mount the file system: mount /mntpnt
File system table As root, edit the /etc/fstab file so that the entry for the file system uses the (Linux and HP-UX only) readthru option, as shown here: /blkdev /mntpnt utdm fstype=fstype,readthru 0 2
Save the file and mount the file system: mount /mntpnt
File system table (AIX only)
As root, edit the /etc/filesystems file so that the options line of the file system’s entry includes the readthru option: options=rw,fstype=fstype,log=/dev/log,readthru
Save the file and mount the file system: mount /mntpnt
122
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
7 Backup and Recovery
This chapter describes backup and recovery of UTDM file systems in the following sections: ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆
What to back up ............................................................................... Recovering a lost file system .......................................................... Recovering an FTP module file system ........................................ EMC NetWorker .............................................................................. NetWorker 64-bit Linux version .................................................... NetBackup software ........................................................................ FSM tools........................................................................................... Snapshots with SnapView .............................................................. Import data from SM to FSM .........................................................
Backup and Recovery
124 128 130 132 136 139 143 147 151
123
Backup and Recovery
What to back up Data that exists solely in UTDM file systems should be backed up regularly to guard against loss caused by inadvertent data deletions or hardware failure. However, migrated data is protected by the back-end system and does not need to be backed up through the file system. When you create a backup strategy for FSM, consider the types of files and data that make up the management domain. These can be generally grouped into the following types: ◆
FSM-related files outside of UTDM file systems
◆
Files and data in UTDM file systems
FSM-related files outside of UTDM file systems At a minimum, the following FSM-related files and directories should be backed up and recovered by using normal backup and recovery methods: ◆
All of /opt/fsm/etc and its subdirectories
◆
/etc/dxuldm.path
◆
/etc/fsm.path
An existing FSM installation can be recovered by reinstalling the software and restoring these files and directories from backup sets.
!
CAUTION If the FSM software must be reinstalled, complete the reinstall before restoring the FSM-related files from backups. This prevents the reinstalled software from overwriting the backup versions of files.
124
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Backup and Recovery
Files and data in UTDM file systems Include the following data in backups of UTDM file systems: ◆
Unmigrated file data.
◆
Stub data for migrated files.
◆
Metadata for all files in the file system.
The backup software described in this chapter automatically includes these entities and excludes unnecessary data from backup sets.
What to back up
125
Backup and Recovery
Backup software The backup software described in Table 36 on page 126 may be used when backing up and recovering data in FSM file systems. Table 36
Supported backup software
Backup software
Usage notes
EMC NetWorker®, release 7.3.2, UNIX version, 64-bit
All FSM-supported 64-bit platforms, except Linux. EMC NetWorker, release 7.3.2, 64-bit Linux version is not FSM-aware. “NetWorker 64-bit Linux version” on page 136 describes how to use it to backup and recover FSM file systems.
EMC NetWorker, release 7.4 SP2 UNIX version, 64-bit
All FSM-supported 64-bit platforms, except Linux. EMC NetWorke release 7.4 sp2, 64-bit Linux version is not FSM-aware. “NetWorker 64-bit Linux version” on page 136 describes how to use it to backup and recover FSM file systems.
EMC NetWorker, release 7.5 SP1 UNIX version, 64-bit
All FSM-supported 64-bit platforms, except Linux.EMC NetWorker, release 7.5 sp1, 64-bit Linux version is not FSM-aware. “NetWorker 64-bit Linux version” on page 136 describes how to use it to backup and recover FSM file systems.
Veritas NetBackup Server version 5.1 All FSM-supported platforms, except SuSE 10. Standard backups only. Veritas NetBackup Server version 6.0 All FSM-supported platforms. Standard backups only. Veritas NetBackup Server version 6.5 All FSM-supported platforms. Standard backups only. FSM backup tools
All FSM-supported platforms.
EMC SnapView™
Snapshots on all FSM-supported versions of HP-UX and Solaris.
!
IMPORTANT Centera data retention can prevent the use of incremental backups to restore some files to a UTDM file system. “Restoring incremental backups” on page 115 explains how to work around this expected limitation.
126
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Backup and Recovery
FSM-aware backup software EMC NetWorker (except for NetWorker, 64-bit Linux version) and FSM backup tools are FSM-aware, and do not cause migrated data to be included in the backup set. Excluding migrated data results in the following advantages: ◆
Shorter backup sessions
◆
Smaller backup sets
◆
Significantly less network traffic
Other backup software As described in Table 36 on page 126, FSM works with the following backup software that is not FSM-aware: ◆
NetWorker, 64-bit Linux version
◆
Veritas NetBackup
When these products are used to back up a UTDM file system, FSM uses its integrated third-party backup support to transparently exclude migrated data from the backup set.
Snapshot software
UTDM file systems can be backed up by using SnapView software. Since only the UTDM file system is included in a snapshot, SnapView does not back up purged data.
Backup software
127
Backup and Recovery
Recovering a lost file system A backup set for a UTDM file system is restored into the existing file system. However, when the file system is lost this cannot be done. The UTDM file system must be replaced. After the file system is replaced, the backup set can be restored into it and the UTDM file system recovered. Note: Since a snapshot is a complete copy of the disk partition, it is not necessary to perform the steps listed in this section when you use a SnapView snapshot to restore a lost file system. The lost UTDM file system is replaced when the snapshot is restored.
To replace a lost UTDM file system: 1. Create a native file system of the same type and same capacity (or greater) as the lost file system, initialize it for UTDM, and mount it on the lost file system’s FSM host. The new file system must be large enough to contain all of the objects that are being restored. To ensure this, the new file system should be at least as large as the lost file system. The installation guides provide more information on these tasks. Note: If the file system’s FSM host has been lost and a new host is being used, the new host must use the name and IP address of the lost host.
2. If any of the following are lost, restore them from normal backup sets: • All of /opt/fsm/etc and its subdirectories • /etc/dxuldm.path • /etc/fsm.path Note: A restored UTDM file system must use the same back-end information as the lost file system. This information is contained in files that exist in /opt/fsm/etc and its subdirectories. By restoring these files you ensure that the recovered UTDM file system can access its back-end data.
128
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Backup and Recovery
3. Complete the appropriate procedure to restore the file system from its backup sets: • “Recovery with NetWorker software” on page 133 • “Recovery with NetBackup software” on page 142 • “Recovery with FSM tools” on page 144
Recovering a lost file system
129
Backup and Recovery
Recovering an FTP module file system A file system that uses the FTP module with the Native FTP protocol cannot create hard links to the data on the back-end system after recovery. This is because the standard FTP command set does not provide the necessary control over the back-end file system. Instead, after the file system has been restored from backup, the file system recovery is completed by mounting the back-end system by using NFS or a similar protocol that provides local disk-like access.
How to recover an FTP module file system To recover a file system that uses the FTP module with the Native FTP protocol: 1. Restore the file system using the appropriate recovery method for the software that created the backup. 2. Using NFS, or a similar protocol, mount the back-end file system on the FSM host. 3. Use Console Client to edit the file system configuration: • Change the Backend Type to NFS. • Change the Target Path to the local mount point of the back-end file system. 4. Log into the FSM host as root. 5. Set the FSM environment. “Setting the FSM environment” on page 177 describes how to do this.
130
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Backup and Recovery
6. Run the dmattrrecoverfs tool: dmattrrecoverfs -p dirpath [-i threads]
where: • dirpath is full path of the managed file system. • threads is the number of parallel forked threads dmattrrecoverfs can start. Appendix A, “Command reference,” provides more information on the dmattrrecoverfs tool. 7. After dmattrrecoverfs has finished the synchronization, use Console Client to edit the file system and return it to its original configuration. The recovery is complete.
Recovering an FTP module file system
131
Backup and Recovery
EMC NetWorker EMC NetWorker is FSM-aware. This sophisticated backup tool detects an FSM file system and performs an intelligent backup of all data and metadata in the file system. These backups do not trigger any staging of data from the back-end system. In the event that data recovery is required, this software restores the backed-up data and initiates a background process that reestablishes all links between restored metadata and data existing on the file system’s back-end system. This background process is called synchronization.
!
IMPORTANT EMC NetWorker, 64-bit Linux version is not FSM-aware. “NetWorker 64-bit Linux version” on page 136 describes how to back up an FSM file system on this platform.
Requirements Path information
Permissions
To use NetWorker for backup and recovery of FSM file systems, the required path information and permissions must be available. To back up an FSM file system by using this software, the file /etc/fsm.path must exist on the file system’s host. This file is automatically created during FSM installation and should not be removed. The nsrexecd daemon controls automated backup and recovery operations of file systems. It is configured with permission to set its user ID as root (suid root). This permission is required for backup and recovery of FSM file systems and should not be changed. Note: Manual NetWorker backup and recovery operations must be performed with root permissions.
Backup with NetWorker software NetWorker backups of FSM file systems are performed just like NetWorker backups of any other file system. The software automatically detects FSM file systems and performs automatic and manual backups without triggering data retrieval from the back-end system. 132
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Backup and Recovery
A full backup set for a FSM file system consists of the following: ◆
Unmigrated file data.
◆
Stub data for migrated files.
◆
Metadata for all files in the file system.
The size of backup sets is significantly reduced. Only the metadata for successfully migrated files is backed up. Note: Migrated file data is recovered by using the metadata backup and the data objects stored on the back-end system.
The .DMATTR file is normally included in the save set created by the NetWorker software. This file can become quite large and does not need to be backed up. It is a volatile file and is completely re-created during recovery. To prevent this file from being backed up: 1. Use the Application Specific Module (ASM) called Skip to create a directive that excludes the .DMATTR file from the save set. 2. During creation of the save set, select the directive in the Directives attribute of the Create Client window. Note: The NetWorker documentation provides more information on creating NetWorker directives and applying directives to save sets.
Recovery with NetWorker software NetWorker recovery of an FSM file system is performed just like NetWorker recovery of any other file system. The entire file system, or just one or more files, may be recovered. When the software restores file data to an FSM file system, the file system’s metadata is automatically synchronized with the data on the back-end system. This synchronization process ensures full access to files that are purged at the time of the backup.
EMC NetWorker
133
Backup and Recovery
The following steps describe the synchronization process: 1. A partial or complete recovery of a FSM file system is performed. After the recovery, the following data is immediately available: • Unmigrated file data • Stub data • Metadata Note: Migrated data becomes available as it is synchronized by the cron job described in step 3.
2. As the NetWorker software performs the file system recovery, it automatically lists each recovered file’s pathname in a file called rec.date.pid, where: • date is the current date. • pid is the process ID associated with the NetWorker client process. The rec.date.pid file exists on the FSM host system in the /opt/fsm/adm/recdir directory. 3. A cron job periodically checks the rec.date.pid file for recovered files, and synchronizes each file it finds. During installation of FSM, a cron job is installed in root’s crontab. This job checks the rec.date.pid file for recovered file entries. When a recovered file entry is found, a background synchronization process starts to synchronize the file’s metadata and data. After the file is synchronized, the file entry is removed from the rec.date.pid file and the file’s data is accessible through the file system. Note: After a NetWorker recovery, the file system is immediately available to be used. However, the full data for a file that has been migrated is not available until synchronization of the file is complete.
134
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Backup and Recovery
You can shorten the time required to synchronize and access a recovered file by performing a manual synchronization that uses the dxuldmrecover command. Appendix A, “Command reference,” provides more information on this command. Note: Backups must be restored into a file system that uses the same back-end system as the original file system. This provides access to the data stored on the back-end system.
EMC NetWorker
135
Backup and Recovery
NetWorker 64-bit Linux version NetWorker 64-bit Linux version software is not FSM-aware. However, new technology in this release enables FSM to detect that this software is being used and to transparently exclude migrated data from the backup set.
Requirements To use NetWorker 64-bit Linux version software to backup UTDM file systems the following requirements must be met: ◆
The file /opt/fsm/etc/backupcmds must exist.
◆
The file /opt/fsm/etc/restorecmds must exist.
Preparation To prepare to use NetWorker 64-bit Linux version software: 1. Install the NetWorker software. 2. Log in to the FSM host as root. 3. Set the FSM environment. For information on setting the FSM environment, see “Setting the FSM environment” on page 177. 4. Change the working directory: cd /opt/fsm/etc
5. Create the backupcmds file: cp backupcmds.default.Linux-emt64 backupcmds
6. Create the restorecmds file: cp restorecmds.default.Linux-emt64 restorecmds
7. Use the NetWorker Application Specific Module (ASM) "Skip" to create a directive that excludes the .DMATTR file from the save set. 8. During creation of the save set, select the directive in the Directives attribute of the Create Client window.
136
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Backup and Recovery
9. Load backupcmds and restorecmds into the kernel: kill.fsm rc.fsm
The core processes are stopped and restarted with the new information.
Backup After the requirements described in “Requirements” on page 136 and “Preparation” on page 136 are met, a backup of a UTDM file system by using NetWorker software is the same as a backup of any other file system that uses a standard backup. During a backup, FSM creates one or more DMAPI sessions and associates files with those sessions. A cron job is run every 10 minutes to start the utility utdm_backupcleanup to clean up those sessions. This cron job is installed in root’s crontab when FSM is installed. This activity is transparent to file system operations. Backup sessions may be viewed by using the FSM tool prtdmsession. The output for a backup session takes the following form: BACKUP filesystem PID ParentPID NWCommand
where: ◆
filesystem is the name of the UTDM file system.
◆
PID is the process ID for the session.
◆
ParentPID is the parent process ID.
◆
NWCommand is the name of the NetWorker command.
For more information about using prtdmsession, see “prtdmsession” on page 188.
Recovery After the requirements described in “Requirements” on page 136 and “Preparation” on page 136 are met, recovery of a UTDM file system by using NetWorker software is the same as the recovery of any other file system. The FSM metadata for each recovered file has the recover bit set. When a file with the recover bit set is accessed it is automatically synchronized.
NetWorker 64-bit Linux version
137
Backup and Recovery
The dmattrrecoverfs tool is also started by an hourly cron job that is installed in root’s crontab. When dmattrrecoverfs is started it scans the file system for files with the recover bit set and synchronizes any it finds. During a recovery, FSM creates one or more DMAPI sessions and associates files with those sessions. A cron job is run every 10 minutes to start the utility utdm_backupcleanup to clean up those sessions. Recovery sessions may be viewed by using the FSM tool prtdmsession. The output for a recovery session takes the following form: RESTORE filesystem PID ParentPID NWCommand
where: ◆
filesystem is the name of the UTDM file system.
◆
PID is the process ID for the session.
◆
ParentPID is the parent process ID.
◆
NWCommand is the name of the NetWorker command.
For more information about using prtdmsession, see “prtdmsession” on page 188.
138
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Backup and Recovery
NetBackup software NetBackup software is not FSM-aware. However, new technology in this release enables FSM to detect that NetBackup software is being used and to transparently exclude migrated data and the .DMATTR file from the backup set. FSM detects a NetBackup backup or restore operation by comparing the NetBackup commands accessing FSM files to command names written in the files backupcmds and restorecmds. FSM installation places default versions of these files in the /opt/fsm/etc directory (where /opt/fsm is the full path to the FSM installation directory). These default versions can be used to create the files backupcmds and restorecmds for use with NetBackup.
NetBackup requirements To use NetBackup software to backup UTDM file systems the following requirements must be met: ◆
One of the following software versions must be used: • The Veritas NetBackup Server version 5.1 (standard backups only; not on SuSE 10) • Veritas NetBackup Server version 6.0 (standard backups only) • Veritas NetBackup Server version 6.5 (standard backups only)
◆
Only standard backups may be run. For example, the FlashBackup package is not supported.
◆
The file /opt/fsm/etc/backupcmds must contain the names of the NetBackup backup commands.
◆
The file /opt/fsm/etc/restorecmds must contain the names of the NetBackup restore commands.
◆
Each NetBackup client machine must have its .DMATTR files listed in the file /opt/openv/netbackup/exclude_list. The .DMATTR file is not needed to restore files from a NetBackup backup.
Preparing to use NetBackup software To prepare to use NetBackup software: 1. Install the NetBackup software. NetBackup software
139
Backup and Recovery
2. Log in to the FSM host as root. 3. Set the FSM environment. For information on setting the FSM environment, see “Setting the FSM environment” on page 177. 4. Add the full path for each .DMATTR file on the FSM host to the NetBackup exclude_list file: echo "/dmattrpath" >> /opt/openv/netbackup/exclude_list
where /dmattrpath is the full path to one of the .DMATTR files on the FSM host. Repeat this command for each .DMATTR file on the FSM host. Example 5 on page 140 provides more information. 5. Change the working directory: cd /opt/fsm/etc
6. Copy backupcmds.default to backupcmds: cp backupcmds.default backupcmds
7. Copy the appropriate restorecmds.default file to restorecmds: cp restorecmds.default.platform restorecmds
where platform is one of the following operating systems for the FSM host: • aix_solaris for AIX and Solaris • hp_linux for HP-UX and Linux 8. Restart FSM to load the contents of backupcmds and restorecmds into the kernel: kill.fsm rc.fsm
The core processes are stopped and restarted, and the names of the NetBackup commands are loaded into the kernel. Example 5
Adding files to the exclude_list file
An FSM host has two UTDM file systems with the following pathnames:
140
◆
/mnt/utdm_fs1
◆
/mnt/utdm_fs2
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Backup and Recovery
Example 5 (continued)
Adding files to the exclude_list file
Run the following commands to add the two .DMATTR files to the exclude_list: echo "/mnt/utdm_fs1/.DMATTR" >> /opt/openv/netbackup/exclude_list echo "/mnt/utdm_fs2/.DMATTR" >> /opt/openv/netbackup/exclude_list
User-initiated backup and restore User-initiated backups and restores are those that are started on a client node. These backups and restores do not check the exclude_list file. The .DMATTR files can be manually excluded from these backups. The NetBackup documentation provides more information about user-initiated backups and restores. Note: A user-initiated restore of a .DMATTR file will fail with an error message. If the .DMATTR files are not excluded from a user-initiated backup, they must be excluded from a user-initiated restore. To do this, refer to the NetBackup documentation.
Backup with NetBackup software After the requirements described in “NetBackup requirements” on page 139 and “Preparing to use NetBackup software” on page 139 are met, a backup of a UTDM file system by using NetBackup software is the same as a backup of any other file system that uses a standard backup. During a backup, FSM creates one or more DMAPI sessions and associates files with those sessions. A cron job is run every 10 minutes to start the utility utdm_backupcleanup to clean up those sessions. This cron job is installed in root’s crontab when FSM is installed. This activity is transparent to file system operations. Backup sessions may be viewed by using the FSM tool prtdmsession. The output for a backup session takes the following form: BACKUP filesystem PID ParentPID NBCommand
where: ◆
filesystem is the name of the UTDM file system.
◆
PID is the process ID for the session.
◆
ParentPID is the parent process ID.
NetBackup software
141
Backup and Recovery
◆
NBCommand is the name of the NetBackup command.
For more information about using prtdmsession, see “prtdmsession” on page 188.
Recovery with NetBackup software After the requirements described in “NetBackup requirements” on page 139 and “Preparing to use NetBackup software” on page 139 are met, recovery of a UTDM file system by using NetBackup software is the same as the recovery of any other file system. The FSM metadata for each recovered file has the recover bit set. When a file with the recover bit set is accessed it is automatically synchronized. The dmattrrecoverfs tool is also started by an hourly cron job that is installed in root’s crontab. When dmattrrecoverfs is started it scans the file system for files with the recover bit set and synchronizes any it finds. During a recovery, FSM creates one or more DMAPI sessions and associates files with those sessions. A cron job is run every 10 minutes to start the utility utdm_backupcleanup to clean up those sessions. Recovery sessions may be viewed by using the FSM tool prtdmsession. The output for a recovery session takes the following form: RESTORE filesystem PID ParentPID NBCommand
where: ◆
filesystem is the name of the UTDM file system.
◆
PID is the process ID for the session.
◆
ParentPID is the parent process ID.
◆
NBCommand is the name of the NetBackup command.
For more information about using prtdmsession, see “prtdmsession” on page 188.
142
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Backup and Recovery
FSM tools This release includes tools that perform basic backup and recovery operations for FSM file systems. These tools can be used to: ◆
Back up a file system, without triggering data retrieval.
◆
Partially or completely recover a file system’s data.
◆
Synchronize recovered metadata with data on the back-end system.
The tools are located in the /opt/fsm/adm/bin directory.
Backup with FSM tools To manually back up an FSM file system, use dxuldm_tar. This tool is based on tar and includes the ability to back up data in an FSM file system without triggering data retrieval. When using dxuldm_tar, the .DMATTR file should be excluded from the backup by using the --exclude option. The .DMATTR file is rebuilt during recovery so backing it up is unnecessary. Note: The .DMATTR can become quite large. Excluding it can significantly decrease the time required to back up a file system.
How to back up a file system with FSM tools To manually create a full backup of a file system: 1. Log in as root on the host system. 2. Set the FSM environment. “Setting the FSM environment” on page 177 provides more information. 3. Change the current working directory to the root of the file system that is being backed up.
FSM tools
143
Backup and Recovery
4. Run dxuldm_tar: dxuldm_tar --exclude=.DMATTR -cvf tarfile .
where tarfile is the full pathname of the backup archive file. The period at the end of the command indicates that the backup should include all files in the current working directory and all subdirectories.
Recovery with FSM tools Recovery of a file system consists of the following tasks: ◆
Restore the data to the file system.
◆
Synchronize the file system’s metadata with the data stored on the back-end system.
Table 37 on page 144 shows the task performed by each FSM tool during recovery. Table 37
FSM recovery commands Task
Command
Recovery of a file system, directory, or individual file.
dxuldm_tar
Scan file system for files with the recover bit set and synchronize all of the data for those files with the data on the back-end system.
dmattrrecoverfs
Users may access a file system as soon as dxuldm_tar finishes a recovery. The FSM metadata for each recovered file has the recover bit set. When a file with the recover bit set is accessed it is automatically synchronized. The dmattrrecoverfs tool is also started by an hourly cron job that is installed in root’s crontab. When dmattrrecoverfs is started, it scans the file system for files with the recover bit set and synchronizes any it finds.
144
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Backup and Recovery
How to recover a file system with FSM tools To recover a file system: 1. Log in as root on the host system. 2. Set the FSM environment. “Setting the FSM environment” on page 177 provides more information. 3. Change the current working directory to the root of the file system that is being restored. 4. Run dxuldm_tar: dxuldm_tar -xvf tarfile
where tarfile represents the full path of the file system’s backup file. The dxuldm_tar command extracts the entire backup archive into the current directory, maintaining the original file system structure. The file system is recovered. Data for files is synchronized automatically by dmattrrecoverfs which is started by the script dxuldmhourlycronscript. The script dxuldmhourlycronscript is started by an hourly cron job that is installed in root’s crontab when FSM is installed. Recovered files may be accessed immediately. When a recovered file that has not been synchronized is accessed, it is automatically synchronized. How to recover individual files with FSM commands To recover a file: 1. Log in as root on the host system. 2. Set the FSM environment. For information on setting the FSM environment, see “Setting the FSM environment” on page 177. 3. Change the current working directory to the root of the file system that was backed up.
FSM tools
145
Backup and Recovery
4. Run dxuldm_tar: dxuldm_tar -xvf tarfile rel-filepath
where: • tarfile represents the full path of the backup file. • rel-filepath represents the relative path of the file being restored (relative to the root of the file system that was backed up). The file is recovered and may be accessed immediately. A recovered file that has not been synchronized is automatically synchronized when it is accessed.
146
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Backup and Recovery
Snapshots with SnapView This release provides three tools to support snapshots of UTDM file systems with SnapView. The tools do the following: ◆
Quiesce a mounted UTDM file system to permit a snapshot to be created.
◆
Resume the UTDM file system activity after a snapshot is created.
◆
Set the recover bit on all migrated files in a restored snapshot.
For information on creating and restoring a snapshot, see your SnapView documentation. Note: The ability to produce SnapView snapshots of UTDM file systems is not available on all of the operating systems that are supported by this release. The release notes provide an up-to-date list of the operating systems on which it is available.
“Setting the FSM environment” on page 177 describes how to set the FSM environment.
Quiescing a UTDM file system The FSM tool, utdm_fs_freeze, quiesces mounted UTDM file systems for a specified number of seconds. This causes the following: ◆
Any thread waiting to access the file system is suspended.
◆
All current operations are concluded.
◆
A flush of any dirty buffers in the file system.
◆
A stable image on the disk.
Use utdm_fs_freeze to prepare a mounted file system for a snapshot. This tool is not required when the file system is not mounted. To quiesce a UTDM file system: 1. Log in to the file system’s host as root. 2. Set the FSM environment.
Snapshots with SnapView
147
Backup and Recovery
!
IMPORTANT A snapshot must be completed before the file system resumes activity or it may be corrupt. Use the -t option for utdm_fs_freeze to specify sufficient time to complete the snapshot. 3. Run utdm_fs_freeze: utdm_fs_freeze -t sec mountpoint
where: • sec is the number of seconds to quiesce the file system. This can be from 1 to the maximum time-out value. The default for the maximum time-out value is 300 seconds. Note: The maximum time-out value can be changed by using the FSM tool, set_freeze_timeout.
• mountpoint is the full path of the directory on which the file system is mounted. The SnapView snapshot may now be created. File system activity automatically resumes after the number of seconds specified by the -t option. To resume file system activity before that time, use utdm_fs_thaw. Determining maximum time-out value To determine the current maximum time-out value: 1. Log in to the file system’s host as root. 2. Set the FSM environment. 3. Run get_freeze_timeout: get_freeze_timeout
The current maximum time-out value is displayed. Setting the maximum time-out value To set the maximum time-out value: 1. Log in to the file system’s host as root. 2. Set the FSM environment.
148
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Backup and Recovery
3. Run set_freeze_timeout: set_freeze_timeout newtimeout
where newtimeout is the number of seconds of the new time-out value. The maximum time-out value is changed to the new value.
Resuming UTDM file system activity The FSM tool, utdm_fs_thaw, resumes activity for UTDM file systems that have been quiesced by using utdm_fs_freeze. After utdm_fs_thaw is run, all suspended threads are allowed to complete. This tool is used after a snapshot is created by using SnapView. To resume activity on a UTDM file system: 1. Log in to the file system’s host as root. 2. Set the FSM environment. For information on setting the FSM environment, see “Setting the FSM environment” on page 177. 3. Run utdm_fs_thaw: utdm_fs_thaw mountpoint
where mountpoint is the full path of the directory on which the file system is mounted.
Recovering a UTDM file system snapshot The FSM tool, utdm_recdmattrf, processes UTDM file systems that are restored from SnapView snapshots. The tool sets the recover bit on all migrated files in the recovered file system. When these files are accessed the core processes create a new namespace for the files and resets the recover bits. A new namespace for migrated files prevents conflicts that would occur if both the original and the restored file system accessed the back-end data from the same namespace. To recover a UTDM file system snapshot: 1. Log in to the file system’s host as root. 2. Set the FSM environment.
Snapshots with SnapView
149
Backup and Recovery
3. Mount the file system that was recovered from a snapshot by using the file system’s native type, as shown in Table 38 on page 150. The variables for the mount procedures are as follows: • fstype is the native type of the file system (not utdm). • blkdev is the full path to the file system’s block special device. • mntpnt is the full path to the file system’s mount point. Table 38
Mounting a file system for utdm_recdmattrf Operating system
Procedure
All operating systems, except Linux and AIX
As root, run the following mount command:
Linux
As root, run the following mount command:
mount-F fstype /blkdev /mntpnt
mount-t fstype /blkdev /mntpnt
AIX
As root, run the following mount command: mount-V fstype /blkdev /mntpnt
4. Run utdm_recdmattrf: utdm_recdmattrf /dmattrpath
where /dmattrpath is the full path to the .DMATTR file of the restored snapshot. When the command prompt is returned (exit status 0) the restored UTDM file system snapshot is recovered. 5. Unmount the file system: umount /mntpnt
6. Mount the file system as type utdm. The appropriate operating system version of the installation guide provides more information on mounting the file system.
150
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Backup and Recovery
Import data from SM to FSM EMC DIskXtender for Unix/Linux Store Manager which was a component of DiskXtender, provided a central data storage repository that utilizes a unified namespace and delivers shared file access using FTP or NFS. To import data from SM to FSM: 1. Install FSM with Centera as a backend. For more information on installing FSM, see EMC DiskXtender File System Manager for UNIX/Linux, Release 3.5 Installation Guides. 2. Mount SM through NFS using the following command: mount -o intr,port=2050 137.69.XX.XX:/DISKXTENDER/ /tmp/IMPORT_DIR
3. To ensure the SM file system is mounted, type: mount | grep DISKXTENDER
The following output is displayed: 137.69.XX.XXX:/DISKXTENDER/ on /tmp/IMPORT_DIR type nfs (rw,intr,port=2050,addr=137.69.44.XXX)
4. In the FSM system with SM mounted through NFS, change the directory to point to the SM mount point and perform the following to populate the UTDM file system with SM data: a. Run the dxprtstat utility for all the files in SM. The output generated to a file in the local file system displays the DXUL-SM object attributes information that are necessary to load objects into a FSM file system. Example: dxprtstat -i {} \; > /tmp/dxprtstat.out
The output looks like: PATH: /tmp/IMPORT_DIR/temp/100_100_R_001 TYPE: FILE MODE: 0750 SIZE: 1 UID: 0 GID: 0 ATIME: 0x494128d4 MTIME: 0x494128d4 CTIME: 0x494128d4 Import data from SM to FSM
151
Backup and Recovery
FILETARGET: ./c5/93/abcdefab00004f5254f88da156fdd293@54f88da1@I MPORT PATH: /tmp/IMPORT_DIR/temp/100_100_R_002
b. Run the dxhardlink utility to create hard links in the SM directory HARDLINKSDIR for SM data that are imported to FSM. Example: dxhardlink -t /tmp/IMPORT_DIR/HARDLINKDIR < /tmp/dxprtstat.out
c. Run the dxbuildtar utility to generate a .tar file in the local file system . Example: dxbuildtar -f /tmp/dxbuildtar.tar -p /tmp/IMPORT_DIR/ < /tmp/dxprtstat.out
The output of the .tar file looks like the SM file system directory structure as below: … -rwxr-x--- root/root ./temp/100_100_R_004 -rwxr-x--- root/root ./temp/100_100_R_005 -rwxr-x--- root/root ./temp/100_100_R_006 -rwxr-x--- root/root ./temp/100_100_R_007 -rwxr-x--- root/root ./temp/100_100_R_008
1024 2008-12-11 14:51:00 1024 2008-12-11 14:51:00 1024 2008-12-11 14:51:00 1024 2008-12-11 14:51:00 1024 2008-12-11 14:51:00
d. Change the working directory to FSM File directory. Run the dxdmimport utility to load objects to an FSM file system from a tar file. /opt/fsm/adm/smtodm/dxdmimport -xUvf /tmp/dxbuildtar.tar
Imported files in FSM will not look the same as normal files in FSM. A getfileattr on a file that is just imported will look like: getfileattr ./100_100_R_001 dt_path: /mnt/utdm/2SMRESTORZE/temp/./100_100_R_001 dt_handle: 32-F9C57356476C4F4823380000000000007AC1C38000000000 0102000000000000
152
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Backup and Recovery
dt_ino: 14371 dt_mode: FILE 750 dt_nlink: 1 dt_uid: 0 dt_gid: 0 dt_size: 1 dt_atime: 1425659730 Fri 06 Mar 2015 04:35:30 PM UTC dt_mtime: 1229007060 Thu 11 Dec 2008 02:51:00 PM UTC dt_ctime: 1425659730 Fri 06 Mar 2015 04:35:30 PM UTC dt_blksize: 4096 dt_blocks: 0 dmattr0: usiversn (UVDM 1.1) dmattr1: usiismig (1) dmattr2: usiispur (1) dmattr3: usifilev (1) dmattr4: usiopath (./c5/93/abcdefab00004f5254f88da156fdd293@54f88da1@ IMPORT) dmattr5: usimtime (494128d4) dmattr6: usimigct (494128d4) dmattr7: usipurtm (494128d4) numregions: 1 regoff0: 0 regsz0: 0 regflg0: rwt ApplFlags: MIGRATED|PURGED|IMPORT_PENDING RevPath: 2SMRESTORZE/temp/100_100_R_001
5. Stop the FSM services and edit the DXULDM.conf file. For FTP, add the following: IMP_0_MODULE:1:FTP IMP_0_FTPHOST:1:137.69.44.XXX IMP_0_FTPPORT:1:1021 IMP_0_FTPUSER:1:root IMP_0_FTPPASSWORD:1:foobar IMP_0_FTPDIRECT:1:On IMP_0_TARGET:1:
For NFS, add the following: IMP_0_MODULE:1:NFS IMP_0_TARGET:
6. Restart the FSM services. The file system is imported. Attempt to read data with import_pending flag will automatically move the whole data from SM to FSM by dxuldmimportfs which is started by the script dxuldmhourlycronscript. The script dxuldmhourlycronscript is started by an hourly cron job that is installed in root’s crontab when FSM is installed. Import data from SM to FSM
153
Backup and Recovery
Note: Do not use the FSM GUI for any FSM configuration changes until all SM data is migrated to FSM or Centera. FSM GUI does not understand the new parameters entered in the DXULDM.conf file and any trigger to rewrite or update will result in loss of information in the DXULDM.conf file.
154
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
8 Maintenance
This chapter covers the following topics: ◆ ◆ ◆
Processes............................................................................................ 156 File administration........................................................................... 164 Periodic back-end system cleanup ................................................ 167
Maintenance
155
Maintenance
Processes FSM consists of the following general categories of processes:
Core processes
◆
“Core processes” on page 156
◆
“Console processes” on page 162
The FSM core processes do the following: ◆
Monitor the status of file systems.
◆
Migrate file data to the back-end system.
◆
Purge file data from file systems.
◆
Retrieve data from the back-end system.
◆
Provide file system information to Console Agent.
◆
Perform task requested by Console Agent.
For each managed file system the following core processes are started: ◆
Two migpurged processes
◆
Two uvdmigd processes
The migpurged process A migpurged process monitors each file system to determine when to do the following: ◆
Migrate file data
◆
Purge file data
The parameters of the monitoring that migpurged performs vary based on the file system’s configuration. Two migpurged processes start for each file system:
156
◆
The first process is a persistence daemon, and restarts the primary daemon as needed.
◆
The second process is the primary daemon and performs the file system management tasks.
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Maintenance
The uvdmigd process
The uvdmigd process ensures that migrated and purged file data is available as needed. When a request is made for file data that resides on the EMC Centera, the uvdmigd process: ◆
Calls the migin process to retrieve the data.
◆
Monitors the retrieval.
◆
Notifies the system that the data is available.
Two uvdmigd processes are started for each file system: ◆
The first process is a persistence daemon, and restarts the primary daemon as needed.
◆
The second process is the primary daemon and performs the file system management tasks.
During a data retrieval, the requesting process waits until sufficient data is retrieved for its operation to complete. Other processes seeking access to the same data also wait. Processes initiating operations that require data already in the file system proceed without delay. Deleted Files The uvdmigd process monitors the file system for files which are removed by users. When a migrated file is deleted from the file system, uvdmigd calls the migdestroy process. The migdestroy process sets a flag on the back-end system to indicate that the file has been deleted from the file system. The indicator that is used varies depending upon the type of back-end system. EMC Centera When a file is deleted from a file system that uses an EMC Centera back-end system, a flag is set in each of the C-Clips for the file. The flag is used to indicate that the file has been deleted from the file system. The file’s data and C-Clips remain on the EMC Centera until removed by an administrator. “EMC Centera” on page 168 provides more information. FTP or NFS module back-end system When a file is deleted from a file system that uses either an FTP or NFS module back-end system, the migdestroy process renames the file on the back-end system by appending the extension .deleted to the filename.
Processes
157
Maintenance
These renamed files remain on the back-end system until removed by an administrator. “Cleanup on FTP or NFS module back-end systems” on page 170 provides more information. Starting and stopping core processes FSM is designed to start when the host system boots up, and to stop when the host system shuts down. During installation, the rc.fsm script is placed in the host system’s startup directory to provide for an orderly startup when the host is booted. The rc.fsm script can also be used to start the FSM processes manually as described in “Manually starting core processes” on page 158. Commands contained in the rc.fsm script perform initialization tasks and start the processes. The script does the following: ◆
Parses the configuration file to determine the following: • File systems to manage. • Configuration settings for each file system.
◆
Starts the core processes for each file system.
◆
Starts Console Agent and, if present on the host, Console Server.
During installation the kill.fsm script is placed in the host system’s shutdown directory to provide for an orderly shutdown of the core processes when the host is shutdown. The kill.fsm script can also be used to stop the FSM processes manually as described in “Stopping FSM processes” on page 159. Setting the FSM environment Most FSM command-line tasks require that you set the FSM environment before running the command. “Setting the FSM environment” on page 177 provides more information. Manually starting core processes An administrator may start the core processes manually by using the rc.fsm script. When installed as recommended, FSM starts automatically each time the system boots up. This is the preferred method of starting the core processes. It ensures that the environment is set up properly, and the required processes are started for every file system.
158
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Maintenance
!
IMPORTANT All UTDM file systems must be mounted before running rc.fsm. Normally these file systems are mounted using the file system table file during boot up. How to manually start the core processes To manually start the core processes: 1. Log in as root on the host system. 2. Set the FSM environment. To set the environment, refer to “Setting the FSM environment” on page 177. 3. Run the rc.fsm script: /opt/fsm/etc/rc.fsm
The core processes start for each file system. Stopping FSM processes When installed as recommended, FSM stops automatically each time the system shuts down. This is the preferred method of stopping the FSM processes. It ensures that for every file system that is being managed, all pending tasks are completed before shutdown occurs. Manually stopping processes An administrator can manually stop the core processes by using the kill.fsm script. Effect of manually stopping processes
The core processes should rarely be stopped. Stopping the processes results in an inability to migrate data to, retrieve data from, and manage data on the back-end system. Requests that involve any of these tasks remain pending until the required process is restarted. Once the process is restarted, pending requests complete. Note: When it is necessary to stop the core processes, such as when a file system must be unmounted, use the kill.fsm script. This script terminates all core processes for all file systems. Restart those processes as soon as possible, either by rebooting the system or by manually running the rc.fsm script. How to manually stop the core processes
To manually stop the core processes:
Processes
159
Maintenance
1. Log in as root on the host system. 2. Set the FSM environment. To set the environment, refer to “Setting the FSM environment” on page 177. 3. Run the kill.fsm script: /opt/fsm/etc/kill.fsm
All core processes on the host stop. Note: The kill.fsm script does not unmount file systems. If the core processes for a file system will not be restarted, be sure to manually unmount that file system.
Monitoring core processes The performance of each process can be monitored through its log. All logs for core processes are located in /opt/fsm/adm/log. When a log reaches 10 MB it is archived by appending a timestamp to its file name and creating a new log. The timestamp that is appended to an archived log is in the format: .YYYYMMDDhhmmss where:
160
◆
YYYY is the four-digit year.
◆
MM is the two-digit month.
◆
DD is the two-digit day.
◆
hh is the two-digit hour.
◆
mm is the two-digit minute.
◆
ss is the two-digit second.
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Maintenance
Each file system has the following logs for its core processes: ◆
migpurged_fsid.t (current log)
◆
uvdmigd_fsid.t (current log)
◆
migpurged_fsid.t.timestamp (archived log)
◆
uvdmigd_fsid.t.timestamp (archived log)
where: ◆
fsid is the file system ID
◆
.timestamp is the timestamp of an archived log.
The logs are recorded in a tokenized text format to reduce their size. To view a log’s contents in a meaningful fashion, use the log viewing command, dx_read_log. Note: An administrator may provide localized, custom log results by using the dx_read_log command. The command replaces tokens in a log with messages from the Common Message File (CMF) and the optional Site Message File (SMF). The SMF, if it exists, is searched first. The messages in the SMF file can be changed to provide users with custom log results.
How to view a log To use dx_read_log to view a log: 1. Log in as root on the host system. 2. Set the FSM environment. To set the environment, refer to “Setting the FSM environment” on page 177. 3. Run dx_read_log: dx_read_log option -l /opt/fsm/adm/log/log-file
where: • option is a viewing option. • log-file is the name of the log. Note: A detailed description of the options available for dx_read_log is available in “dx_read_log” on page 179.
Processes
161
Maintenance
Console processes
The FSM Console processes consist of the Console Server and the Console Agent. The Console Server runs on one host in the management domain. It does the following: ◆
Maintains user accounts and roles.
◆
Negotiates authentication between Console Client and Console Agents.
The Console Agent runs on each FSM host. It provides the following: ◆
Initial authentication
◆
File system monitoring
◆
File system configuration
Starting and stopping Console processes The Console processes start and stop automatically when the host’s core processes are started and stopped. You should not start or stop the Console processes any other way unless you are instructed to by Customer Support. Monitoring Console processes The performance of all Console Server and Console Agent processes is recorded in a log on the computer that hosts the process. The logs are located in the following directory: /opt/fsm/logs
Console Server log The Console Server log is named: FSMConsoleServer.log
This log grows until it reaches the number of kilobytes specified by the value of the fsm.log.size parameter in the following file: /opt/fsm/etc/FsmConsoleServer.properties
This value can be changed by editing the file to uncomment the parameter and change its value. The default is 5,000,000 KB. When the maximum size is reached the log is archived and a new log is started. Archived logs have a timestamp appended to their filename. To reclaim space you can manually remove the archived logs. 162
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Maintenance
Console Agent log The Console Agent log is named: FSMConsoleAgent.log
This log grows until it reaches the number of kilobytes specified by the value of the fsm.log.size parameter in the following file: /opt/fsm/etc/FsmConsoleAgent.properties
This value can be changed by editing the file to uncomment the parameter and change its value. The default is 5,000,000 KB. When the maximum size is reached the log is archived and a new log is started. Archived logs have a timestamp appended to their filename. To reclaim space you can manually remove the archived logs. Command output logs Both Console processes produce a second set of logs. These are the command output logs, which contain the messages that are normally displayed on a console. The Console Server command output log is named: FSMConsoleServer_log
The Console Agent command output log is named: FSMConsoleAgent_log
These logs grow at a much slower rate, but their size is controlled by the same parameter as the Console Server log and the Console Agent log. To reclaim space you can manually remove the archived versions of these logs.
Processes
163
Maintenance
File administration FSM provides tools to perform administrative tasks on individual files.
Manual migration, purge, and retrieval An administrator can manually migrate, purge, and retrieve individual files as needed. How to manually migrate a file To manually migrate a file: 1. Log in as root on the host system. 2. Set the FSM environment. To set the environment, refer to “Setting the FSM environment” on page 177. 3. Change the working directory to the directory that contains the file. 4. Run migout: migout filename
where filename is the name of the file. Upon successful migration, migout displays the message: Successfully migrated file filename Note: If this message does not appear, the file was not migrated.
How to manually purge files To manually purge files: 1. Log in as root on the host system. 2. Set the FSM environment. To set the environment, refer to “Setting the FSM environment” on page 177. 3. Change the working directory to the directory that contains the file. 4. Run purge: 164
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Maintenance
purge filename
where filename is the name of the file. Upon successful purging, purge displays the message: Successfully purged file filename. Note: If this message does not appear, the file was not purged. Manual purging fails if the file’s Purge attribute has been set to NEVER, or if the file has not been migrated. For files which have not been migrated, the purge command displays the message: "File filename has not been migrated."
The successful completion of the purge operation can be verified by using the list command, ls -ls, on the file both before and after purging. The number of data blocks occupied by the file on the file system is less after successful purging. Note: Only files that are eligible for purging can be manually purged. To be eligible for purging, a file must have been successfully migrated and must be unchanged since the migration. As expected, an attempt to purge a file that contains data that has not been migrated will fail.
How to manually retrieve files The uvdmigd process automatically retrieves a purged file when needed. Very large files can cause a noticeable time delay while the file data is retrieved. To manage this time delay, the large file can be retrieved in advance. To manually retrieve files: 1. Log in as root on the host system. 2. Set the FSM environment. To set the environment, refer to “Setting the FSM environment” on page 177. 3. Change the working directory to the directory containing the file.
File administration
165
Maintenance
4. Run the list command, ls -ls: ls -ls filename
where filename is the name of the file. The number of blocks that the file occupies in the file system are displayed. 5. Run migin: migin filename
where filename is the name of the file. 6. Run the list command: ls -ls filename
where filename is the name of the file. The number of blocks occupied by the file on the file system increases after the file’s data is successfully retrieved.
166
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Maintenance
Periodic back-end system cleanup Files which have been deleted from a file system are not automatically removed from the back-end system. This permits the recovery of inadvertently removed files. Chapter 7, “Backup and Recovery,” provides more information on recovery. When a user deletes or removes a file from a file system, the following occurs: 1. The uvdmigd process reads the file’s DMAPI attributes to determine if the file has been migrated. 2. If the file has been migrated, uvdmigd calls the migdestroy process. 3. The migdestroy process sets a flag on the back-end system to indicate that the file has been deleted from the file system. The indicator used varies depending upon the type of back-end system. Periodically, an administrator should make available for purging the deleted file data that is stored on the back-end system. This is called cleanup. For the following back-end module types, cleanup may be scheduled to run automatically by using Console Client: ◆
Centera
◆
NFS
Chapter 5, “Scheduled Jobs,” and the online help describe this. Back-end systems that use the FTP module require manual cleanup. On all back-end system types you may perform manual cleanup. This is described in the following sections:
!
◆
“EMC Centera” on page 168
◆
“Cleanup on FTP or NFS module back-end systems” on page 170
IMPORTANT Only perform cleanup of data on the back-end system when you are sure that it is not needed. Purging data from the back-end system permanently removes it. This data cannot be recovered.
Periodic back-end system cleanup
167
Maintenance
EMC Centera
For file systems that use the Centera module, cleanup on the EMC Centera is performed from the FSM host by using the dxuldmdelclips command to identify and mark a range of files. Cleanup is performed simultaneously for all Centera module file systems on a host. The dxuldmdelclips command must have query permission on the EMC Centera to perform its tasks. If the default permissions are changed to deny query permission for anonymous profiles, this permission must be provided through a PEA file. The file system’s PEA file is identified through its Centera Profile Path value. “Centera module preliminary tasks” on page 45 provides more information.
How to clean up an EMC Centera To clean up an EMC Centera: 1. Log in as root on the host system. 2. Set the FSM environment. To set the environment, refer to “Setting the FSM environment” on page 177. 3. Run dxuldmdelclips: dxuldmdelclips -a emc-address[?/PEA-file] -v
where: • emc-address is one of the IP addresses listed in the Connect Addresses value. “Centera module” on page 64 provides more information. • /PEA-file is the full path to a PEA file which provides read, write, exist, and query permissions. Note: The ?/PEA-file portion of the command is included if authentication is required by the EMC Centera.
For all file systems, the file data that was deleted more than one year before the current date is marked for purging. All C-Clips, linked to the deleted files, are also marked. This procedure is demonstrated in Example 6 on page 169.
168
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Maintenance
Note: To mark the data for a specific range of files follow the procedure described in “dxuldmdelclips” on page 184.
Example 6
Periodic Centera system cleanup
In this example the following parameters are assumed: ◆
The EMC Centera has an IP address of 137.69.44.212
◆
The EMC Centera administrator has provided a PEA file which provides read, write, exist, and query permissions.
◆
The PEA file is located at /etc/rweq.pea.
To clean up all files deleted more than a year ago, set the FSM environment and run the command: dxuldmdelclips -a 137.69.44.212?/etc/rweq.pea -v
Restarting the task
The -v option to dxuldmdelclips causes the command to display status information after every 100 C-Clip deletions. This information may be used to restart the task if it stops prior to completion. To restart an incomplete dxuldmdelclips task: 1. Note the last displayed timestamp from the incomplete dxuldmdelclips run. The timestamp is displayed as seconds since the epoch. 2. Log in as root on the host system. 3. Set the FSM environment. “Setting the FSM environment” on page 177 describes how to do this. 4. Use dxuldmdelclips with the -t option:
dxuldmdelclips -a emc-address?/PEA-file -v -t starttime
where: • emc-address is one of the IP addresses listed in the Connect Addresses value. For more information on this value, refer to “Centera module” on page 64. • /PEA-file is the full path to a PEA file which provides read, write, exist, and query permissions.
Periodic back-end system cleanup
169
Maintenance
• starttime is the last displayed timestamp from the incomplete dxuldmdelclips run. The cleanup task starts and marks data as available for purging. Only data that was deleted during the time range specified is marked as available for purging. The default values for dxuldmdelclips marks files that were deleted in the time range between the following two dates: • The time specified by starttime. • The date that is one year before when the command is run.
Cleanup on FTP or NFS module back-end systems Cleanup for back-end systems that use the NFS or FTP modules is performed on the back-end host. This cleanup is performed on a directory by directory basis. How to clean up a back-end system that uses either the FTP or NFS module To clean up a back-end system that uses either the FTP or NFS module: 1. Log in as root on the back-end host. 2. Change the working directory to the directory that contains the deleted file data. 3. Run find with the -atime option: find . -name *.deleted -atime +n -print > list-file &
where: • n is an integer that represents the number of days since last access. • list-file is the name of a temporary file to hold the results of the find command. 4. Review list-file. If it includes files which should be kept repeat the command with a greater value for the number of days since last access. 5. If every file listed in list-file can be removed, type the find command with the -exec switch that calls the rm command: find . -name *.deleted -atime +n -exec rm {} \; &
170
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Maintenance
where n is an integer representing the number of days since last access. The files listed in list-file are removed.
Periodic back-end system cleanup
171
Maintenance
172
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
A
Invisible Body Tag
Command reference
This appendix covers the following topics: ◆ ◆
Command quick reference.............................................................. 174 Command usage .............................................................................. 177
Command reference
173
Command reference
Command quick reference The tables in this appendix list the commands used with this software release: ◆
Administrative Commands, refer to Table 39.
◆
File Management Commands, refer to Table 40 on page 175.
◆
Backup and Restore Commands, refer to Table 41 on page 175.
“Command usage” on page 177 provides detailed information about each command. Except where otherwise specified, all commands are located in /opt/fsm/adm/bin. and the commands are described in man pages located in /opt/fsm/man, where /opt/fsm is the full path to the FSM installation directory. To access the man pages, set the FSM environment, and type man followed by the name of one of the commands. For information on setting the FSM environment, refer to “Setting the FSM environment” on page 177. Table 39
174
Administrative commands (page 1 of 2)
Command
Description
deldmsession
Deletes abandoned DMAPI sessions.
dx_read_log
Preprocesses logs by replacing tokens with messages, and displays the result.
dxuldmcenteraping
Tests connectivity with network interfaces on an EMC Centera and displays the result. When a connection is successfully made, the hardware and firmware values for the EMC Centera are displayed.
dxuldmcheckconf
Checks the configuration file for inconsistent and insufficient entries.
dxuldmclip
Displays the C-Clip associated with a given address and C-Clip ID.
dxuldmdelclips
Searches a range of C-Clips and marks those for deleted files as available for purging from the EMC Centera.
expand_dmattrf
Expands a file system’s .DMATTR file.
getdmattr
Reports the value of a specific FSM attribute for one or more files.
getfileattr
Reports all FSM and DMAPI debugging information for a file.
prtdmsession
Prints information that describes all the DMAPI sessions that run on the local computer.
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Command reference
Administrative commands (page 2 of 2)
Table 39
Command
Description
dxstat
Prints DXUL-FSM migrate, purge, stage, delete, recover and import statistics.
dxprtstat
Generate output that describes the attribute information necessary to load a new object into an FSM file system that is a mirror image of what the object looks like through the DXUL-SM interface.
dxhardlink
Generates hardlinks within the DXUL-SM system.
dxdmimport
Loads objects into an FSM managed file system. File management commands
Table 40
Table 41
Command
Description
getimplist
Retrieves a list of files in an FSM managed file system which need to be imported.
getmiglist
Displays a list of files that need to be migrated.
getpurgelist
Displays a list of files that need to be purged.
migin
Retrieves data from the back-end system.
migout
Migrates data to the back-end system.
migpurgestate
Displays the migration and purge status of files.
purge
Purges file data from a file system.
Backup and restore commands (page 1 of 2) Command
Description
dxbuildtar
Generates a tar-formatted file that is used to load objects into a UTDM file system.
dxcliplink
Create a link to a C-Clip and print out the C-Clip attributes required to create a UTDM file system’s stub file linked to the data represented by the C-Clip.
dxdmload
Load objects into a UTDM file system from a file generated by dxbuildtar.
dxuldm_tar
Backs up and restores file systems, without triggering a retrieval of migrated and purged data.
dmattrrecoverfs
Synchronizes all file system metadata with data stored on the back-end system.
175
Command reference
Table 41
176
Backup and restore commands (page 2 of 2) Command
Description
get_freeze_timeout
Displays the current maximum time a UTDM file system can be frozen.
getrecoverlist
Returns a list of files that need to be recovered into a UTDM file system.
set_freeze_timeout
Sets the maximum time a UTDM file system can be frozen.
utdm_fs_freeze
Temporarily suspends access to a UTDM file system.
utdm_fs_thaw
Permits access to a UTDM file system that was put in a suspended state by utdm_fs_freeze.
utdm_recdmattrf
Finalizes recovery of a UTDM file system from a snapshot.
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Command reference
Command usage The commands are described in the following sections: ◆
“Administrative commands” on page 178
◆
“File management commands” on page 191
◆
“Backup and recovery commands” on page 196
Setting the DMAP_ROOT_PATH Before using any of the FSM commands, the environment variable DMAP_ROOT_PATH must be set to the full path of the FSM installation directory. The FSM installation directory is the top-level FSM directory (the directory which contains the adm, bin, etc, and man subdirectories). The default path to this directory is /opt/fsm. The DMAP_ROOT_PATH is set by setting the FSM environment, refer to “Setting the FSM environment” on page 177.
Setting the FSM environment Most FSM command-line tasks require that you set the FSM environment before you run the command. To set the FSM environment, type the correct source command for your current shell. ◆
C shell (csh) and Turbo C shell (tcsh), type: source /opt/fsm/etc/fsm.login
where /opt/fsm is the full path to the FSM installation directory. ◆
Bourne shell (sh), Bourne Again shell (bash), and Korn shell (ksh), type: . /opt/fsm/etc/fsm.profile
where /opt/fsm is the full path to the FSM installation directory.
177
Command reference
Administrative commands The administrative commands are described in the following sections:
deldmsession
◆
“deldmsession” on page 178
◆
“dx_read_log” on page 179
◆
“dxuldmcenteraping” on page 182
◆
“dxuldmcheckconf” on page 183
◆
“dxuldmclip” on page 183
◆
“dxuldmdelclips” on page 184
◆
“expand_dmattrf” on page 186
◆
“getdmattr” on page 187
◆
“getfileattr” on page 188
◆
“prtdmsession” on page 188
◆
“dxstat” on page 188
◆
“dxprtstat” on page 189
◆
“dxhardlink” on page 190
◆
“dxdmimport” on page 190
Use the deldmsession command to remove abandoned DMAPI sessions. DMAPI sessions become abandoned due to unexpected termination of a process. These sessions normally cause no problems and are removed when the system is rebooted. However, in some instances, a file system cannot be successfully unmounted if an abandoned DMAPI session is left active. When using this command, avoid accidental removal of active DMAPI sessions. If active sessions are removed, FSM processes terminate and file system activity is blocked until those processes are restarted. Syntax deldmsession [[-s session-string]|[-i session-id]] -n
Options The options for deldmsession are described in Table 42 on page 179.
178
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Command reference
Table 42
Options for deldmsession
Option
Description
-s session-string
The -s option takes the argument session-string which is the label of the session to be removed.
-i session-id
The -i option takes the argument session-id which is the integer identification number for the session being removed.
-n
The -n option prevents deldmsession from responding to system events.
dx_read_log
The dx_read_log command starts a log preprocessor that expands token logs into plain text format. It replaces the tokens with messages from the Common Message File (CMF) and the optional Site Message File (SMF). If the SMF exists, it is searched first. After all replacements are made, the requested portion of the processed log is displayed. Syntax
dx_read_log -l /log-file [-s [-|=|+] severity] [-T begintime] [-E endtime] [-d begindate] [-D enddate] [-p] [-h]
Options The options for dx_read_log are described in Table 43 on page 180.
179
Command reference
Table 43
180
Options for dx_read_log (page 1 of 2) Option
Description
-l /log-file
The -l option takes the argument /log-file, that represents the full path to the log to be read. This option is required.
-s severity
The -s option takes the argument severity, which is an integer that represents a severity level, 0-9, and returns messages related to that level. When severity is preceded by: • A minus sign (-), then all messages equal to or less than the level specified by severity are returned. • An equal sign (=), then all messages equal to the level specified by severity are returned. This is the default. • A plus sign (+), then all messages equal to or greater than the level specified by severity are returned. Table 44 on page 181 describes the severity levels.
-b brevity
The -b option takes the argument brevity, which is an integer that represents a brevity level, 0-3, and displays information within that level of detail. Table 45 on page 182 shows the brevity levels that are available.
-t number
The -t option takes the argument number, which is an integer that represents the number of messages to display from the end of the log. This argument is similar to the tail command.
-c number
The -c option takes the argument number, which is an integer that represents the number of messages to display from the end of the log. The display is continuous.
-T begintime
The -T option takes the argument begintime which is a time value, in the format hh:mm:ss. All messages, generated after the specified time, are displayed. When this option is used with the -E option, a time segment can be specified.
-E endtime
The -E option takes the argument endtime which is a time value in the format hh:mm:ss. All messages generated before the specified time are displayed. When this option is used with the -T option, a time segment can be specified.
-d begindate
The -d option takes the argument begindate which is a date value in the format yy/mm/dd. All messages generated after the specified date are displayed. When this option is used with the -D option, a date segment can be specified.
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Command reference
Options for dx_read_log (page 2 of 2)
Table 43
Option
Description
-D enddate
The -D option takes the argument enddate which is a date value in the format yy/mm/dd. All messages generated before the specified date are displayed. When this option is used with the -d option, a date segment can be specified.
-p
The -p option displays the current dx_read_log program parameters. No log messages are returned.
-h
The -h option returns dx_read_log help. No log messages are returned. Severity levels
The -s option accepts an integer argument that is used to filter messages based on a severity level. The severity levels described by each integer argument (0-9) are shown in Table 44 on page 181. Severity levels
Table 44
Level Name
Description
9
Panic
Problems that are not recoverable. The process terminated.
8
Severe
Major problems that require administrative action.
7
Warning
Potential problems detected by system.
6
Recover
Problems encountered by processes, from which the processes successfully recovered.
5
Info
General process information.
4
Debug4
Debug messages with the highest level of detail.
3
Debug3
Debug messages with a level of detail lower than level 4.
2
Debug2
Debug messages with a level of detail lower than level 3.
1
Debug1
Debug messages with a level of detail lower than level 2.
0
Debug0
Debug messages with the lowest level of detail.
181
Command reference
Brevity levels
The -b option takes an integer (0-3) that determines the amount of detail displayed. The brevity levels are shown in Table 45 on page 182. Table 45
dxuldmcenteraping
Brevity levels Level
Description
0
Message arguments only
1
Normal
2
Detailed messages
3
Further explanation
The dxuldmcenteraping command checks connectivity with the network interfaces on a EMC Centera. A EMC Centera does not respond to packets sent by the ping command. When a specified IP address is available, dxuldmcenteraping displays the following for that address: Address: IP-address Info Version: infoversion Capacity: capacity Free Space: free Cluster ID: cluster-id Cluster Name: cluster-name Version: version Replica Address: replica-address
When dxuldmcenteraping encounters an unavailable IP address it displays the following: Unable to open cluster at address (IP-address). Error (FP_NO_POOL_ERR)
Syntax dxuldmcenteraping -a emc-address [emc-address...]
Option The dxuldmcenteraping command takes the option emc-address which is an IP address for a network interface on an EMC Centera. Optionally, additional IP addresses, separated by spaces, may be specified.
182
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Command reference
dxuldmcheckconf
Use the dxuldmcheckconf command to check the configuration file for inconsistencies. The configuration file contains necessary configuration information for all FSM file systems. The dxuldmcheckconf command stops on the first inconsistency it finds and returns the inconsistent attribute’s name, and information about the inconsistency. Optionally, the command can be called through the rc.fsm startup script to prevent FSM from starting with configuration errors. Syntax dxuldmcheckconf [-v]
Option The -v option toggles on verbose mode. In this mode, when no error is detected, a summary of the configuration settings for each file system that is specified in the configuration file is returned. When verbose mode is off, and no error is detected, the command prompt is returned. dxuldmclip
The dxuldmclip command retrieves and displays the information contained in a file’s associated C-Clip. The C-Clips are metadata files located on an EMC Centera back-end system. Example 7 on page 184 shows the C-Clip information displayed by dxuldmclip. Syntax dxuldmclip -a emc-address -c clipid
Options The options for dxuldmclip are described in Table 46 on page 183. Table 46
Options for dxuldmclip
Option
Description
-a emc-address
The -a option takes the argument emc-address which is one of the IP addresses specified for the Connect Addresses attribute.
-c clipid
The -c option takes the argument clipid which is the ID string for a file’s associated C-Clip. To obtain clipid, type the command: getdmattr -a usiopath filepath
where filepath is the full path to the file. The value returned is clipid.
183
Command reference
Example 7
Using dxuldmclip to obtain C-Clip information for a file # getdmattr -a usiopath /utdm.1/file01 usiopath value for file /utdm.1/file01 is: 743NEV737O694e23DFUNDMBPT20 # dxuldmclip -a 137.69.44.247 -c 743NEV737O694e23DFUNDMBPT20 Clip name: FSM_utdm.1 Clip ID: 743NEV737O694e23DFUNDMBPT20 Created on: 2002.12.04 21:22:42 GMT Number of Blobs: 1 Number of Bytes in Blobs: 104857600 Tag: FILE_ATTRIBUTES (META_VERSION) = (Version 1.0) (TARGET) = (utdm.1) (UID) = (0) (GID) = (1) (MIGRATE_TIME) = (1039036962) (PATH_NAME) = (/utdm.1/file01) Tag: VERSION_1 (VERSION_TIME) = (1039036962) (VERSION_SIZE) = (104857600) Tag: OLD_CCLIPS No Attributes for this tag
dxuldmdelclips
The dxuldmdelclips command examines a range of C-Clips for those representing deleted files and marks them as available for purging from the EMC Centera. The C-Clips marked as available for purging are not automatically purged. “Centera module” on page 64 provides more information about purging. Only C-Clips linked to files that have been deleted from the file system are marked by this command. When the C-Clips are marked, the data storage space associated with C-Clip can be reclaimed by the EMC Centera’s purging routine. The FSM file system must have query permission on the EMC Centera to use the dxuldmdelclips command. If the default permissions are changed to deny query permission for anonymous profiles, this permission must be provided through a PEA file. The file system’s PEA file is identified through its Centera Profile Path value. “Centera module preliminary tasks” on page 45 provides more information. “How to clean up an EMC Centera” on page 168 shows an example of the dxuldmdelclips command.
184
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Command reference
Syntax dxuldmdelclips -a emc-address[?/PEA-file] [-t starttime] [-e endtime] [-m age] [-n num-to-del] [-v]
Options The options for dxuldmdelclips are described in Table 47 on page 185. Table 47
Options for dxuldmdelclips
Option
Description
-a emc-address [?/PEA-file]
The -a option takes the argument emc-address which is one of the IP addresses specified for the file system’s Connect Addresses attribute. For more information on the Connect Addresses attribute, refer to “Centera module” on page 64. An optional (but required, in most cases) argument is /PEA-file, the full path to a PEA file. This argument follows a question mark that is appended to the emc-address argument. The PEA file must provide read, write, exist, and query permissions.
-t starttime
The -t option takes the argument starttime which is an integer value that represents a number of seconds since the epoch. The time represented by starttime is used as the starting point for the range of C-Clips evaluated for marking. The default is 0, which means the starting point is the oldest C-Clip for the file system. “Restarting the task” on page 169 provides an example showing how to use this option to restart a dxuldmdelclips run.
-e endtime
The -e option takes the argument endtime which is an integer value that represents a number of seconds since the epoch. The time represented by endtime is used as the ending point for the range of C-Clips evaluated for marking. The default is -1, which means the end point is the most recent C-Clip for the file system.
-m age
The -m option takes the argument age which is an integer value that represents the number of days old a C-Clip must be before it is evaluated for marking. The default value is 365, which means only C-Clips older than one year will be marked as available for purging.
-n num-to-del
The -n option takes the argument num-to-del which is an integer that represents the number of C-Clips that can be marked in the current dxuldmdelclips run. When the range of C-Clips available for marking exceeds num-to-del, the oldest C-Clips are marked first. The default value is -1, which means all eligible C-Clips will be marked.
-v
The -v option toggles on verbose output. When verbose output is on, a message is sent to stdout after every 100 C-Clip deletions, and the final tally of deletions is sent to stdout at the end of the run.
185
Command reference
expand_dmattrf
The expand_dmattrf command is used to make changes to a file system’s .DMATTR file. The following changes can be made by using this tool: ◆
Increase the number of files handled.
◆
Increase the number of attribute blocks handled.
◆
Add space at the end of the file.
◆
Change the file system’s aggregate information.
Before using this command the file system on which the .DMATTR file resides must be mounted using its native format.
!
IMPORTANT Before using this command, make a backup of the original .DMATTR file to permit recovery if errors occur with the new .DMATTR file. Syntax
expand_dmattrf [-f files] [-a blocks] [-z size] [-g agginfo] [-o elements] [-h heads] [-t fs] [-v|-q] path newpath
Options At lease one of the action options must be used: -f, -a, -z, -g, -o, or -h. The file system type, -t, option is only required when the attribute file is converted from a version 1 header format to a version 2 header format. Specifying both the quiet, -q, and verbose, -v, options causes an error. The options for expand_dmattr are described in Table 48 on page 186. Table 48
186
Options for expand_dmattrf (page 1 of 2)
Option
Description
-f files
The -f option takes the argument files which is an integer that represents the new number of supported files.
-a blocks
The -a option takes the argument blocks which is an integer that represents the new number of supported attribute blocks.
-z size
The -z option takes the argument size which is an integer that represents the new size of the .DMATTR file in bytes.
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Command reference
Table 48
Options for expand_dmattrf (page 2 of 2)
Option
Description
-g agginfo
The -g option takes the argument agginfo, which is a colon-separated triple that consists of the following aggregate information values: • Number of aggregates • Number of inodes per aggregate • Initial inode offset for each aggregate For example, the triple 4:4194304:128 sets the new aggregate information to be 4 aggregates, each one with 4,194,304 inodes and an initial 128 inode offset.
-o elements The -o option takes the argument elements which is an integer that represents the new number of hash table elements in the overflow area. -h heads
The -h option takes the argument heads which is an integer that represents the new number of hash table list heads in the overflow area. Increase this number to create more list heads and thereby shorten the lists that are traversed when searching for a hash table element in the overflow area. Note: An increase in this value causes more space to be allocated to the list heads in the .DMATTR file.
-t fs
The -t option takes the argument fs which is the native (underlying) file system type on which the .DMATTR file resides. This option should only be used, and is required, when the attribute file must be converted from a version 1 header format to a version 2 header format.
-v
The -v option turns on verbose mode. This causes the display of messages that describe the changes being made.
-q
The -q option turns on quiet mode. This significantly reduces the display of messages.
path
The variable path represents the full path of the .DMATTR file on the native file system.
newpath
The variable newpath represents the output path, that is, the full path of the .DMATTR file generated by this command.
getdmattr
The getdmattr command displays the value of one of the FSM file attributes for one or more files. Syntax getdmattr -a attribute-name filelist
187
Command reference
Options The options for getdmattr are described in Table 49 on page 188. Table 49
Options for getdmattr
Option
Description
-a attribute-name
The -a option takes the argument attribute-name, which is one of the FSM file attributes.
filelist
The getdmattr command takes the argument filelist, which is the full path description of one or more files. For multiple files, separate each path description by a space.
getfileattr
Use the getfileattr command to display all FSM and DMAPI debugging information for a file. Syntax getfileattr [-u] filepath
Options The options for getfileattr are described in Table 50 on page 188. Table 50
Options for getfileattr
Option
Description
-u
The -u option toggles on streamed output. When this option is toggled on, the output of the command is one continuous, unstructured stream.
filepath
The getfileattr command takes the argument filepath, which is the full path to the file about which information is sought.
prtdmsession
The prtdmsession command displays information for all current DMAPI sessions. Syntax prtdmsession
dxstat
188
Use the dxstat command to print DXUL-FSM migrate, purge, stage, delete, recover and import statistics. It also prints information on activities that are inprogress. If no options are passed, all statistics are printed.
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Command reference
Syntax dxstat [-f pathname] [-m] [-p] [-s] [-d] [-r] [-i]
Options The options for dxstat are described in Table 51 on page 189. Table 51
Options for dxstat
Option
Description
-f pathname
The path to the file system being managed whose statistics are to be printed.
-m
Print migrate statistics.
-p
Print purge statistics.
-s
Print stage statistics.
-d
Print delete statistics.
-r
Print recover statistics.
-i
Print import statistics.
dxprtstat
Use the dxprtstat command to generate output that describes the attribute information necessary to load a new object into an FSM file system that is a mirror image of what the object looks like through the DXUL-SM interface. Syntax dxprtstat [-i] pathname
Options The options for dxprstat are described in Table 52 on page 189.
Table 52
Options for dxprstat
Option
Description
-i
Import files into FSM.
pathname
Pathname to the object.
189
Command reference
Note: dxprtstat should be used in conjunction with the system find command when mirroring objects under a particular directory tree. -i option should be passed when files are being imported from DXUL-SM to FSM.
dxhardlink
The dxhardlink command is used to generate hardlinks within the DXUL-SM system. These hardlinks are generated in such a way that they can be used by the FSM system to access file data. If the dxhardlink command fails part way through for any reason, it can be retried with no ill effects (it simply skips over work already done). The DXUL-SM file system to be loaded with hardlinks must be NFS accessible. Syntax dxhardlink -t target_DXSM_directory [-v]
Options The options for dxhardlink are described in Table 53 on page 190.
Table 53
Options for dxhardlink
Option
Description
-t target_DXSM_directory
The path within the DXUL-SM file system which is a target of the FSM managed file system to be loaded with mirrored objects.
-v
Generate output for every hard link attempt.
dxdmimport
The dxdimport command is used to load objects into an FSM managed file system. These objects should mirror those objects processed during the find/dxprtstat step of the conversion process. Syntax dxdmimport [Same as GNU tar(1)]
Options The only options that should be used are xUf and possibly v. The U option should always be used to guarantee clean loads if objects already exist in the file system.
190
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Command reference
File management commands The file commands are described in the following sections:
getimplist
◆
“getimplist” on page 191
◆
“getmiglist” on page 192
◆
“getpurgelist” on page 193
◆
“migin” on page 195
◆
“migout” on page 195
◆
“migpurgestate” on page 195
◆
“purge” on page 196
The getimplist command is used to retrieve a list of files in an FSM managed file system which need to be imported. This utility is dependent on UTDM specific file system commands that are only available on certain platforms. UTDM allows for migration algorithms that scale. Syntax getimplist -p pathname [-a age] [-c cookie] [-n number] [-s size] [-k]
Options The options for getmiglist are described in Table 55 on page 192. Table 54
Options for getmiglist (page 1 of 2)
Option
Description
-p pathname
The path to the file system being managed
-a age
The age in minutes since the import candidates were created. Any files created within the last "age" in minutes will not be a import candidate. This is an optional parameter and if not supplied all import candidates will be returned.
-c cookie
The -c option takes the argument cookie that identifies a location to start the search. This allows for retrievals of only partial import lists, and the ability to start a new import list where the last import list left off. This will allow for import algorithms to be written which do not require an entire list of all import candidates before import can be started. This is an optional parameter and if not supplied the search will start at the beginning of the file system.
191
Command reference
Table 54
Options for getmiglist (page 2 of 2)
Option
Description
-n number
The -n option takes the argument number which is the maximum number of candidates to return in the import list. If this parameter is not supplied then all import candidates will be returned.
-s size
The -s option takes the argument size which represents the bytes size of a file. Only files whose size is greater than or equal to the size are returned. If size if not supplies then the default size is 0 bytes.
-k
The -k option does not print out the file list. At end of processing just print out the number of files to import plus the number of bytes included in all import candidates.
getmiglist
The getmiglist command retrieves a list of files that need to be migrated. Syntax
getmiglist -p fspath [-a age] [-c cookie] [-n num] [-s size] [-k]
Options The options for getmiglist are described in Table 55 on page 192. Table 55
192
Options for getmiglist (page 1 of 2)
Option
Description
-p fspath
The -p option takes the argument fspath which is the full path of the managed file system. This value is required.
-a age
The -a option takes the argument age which is the age in minutes since a file was last modified. All files modified within the time specified by age are excluded from the list. This is an optional value. The default is 0 which ignores when a file was last modified.
-c cookie
The -c option takes the argument cookie which is an integer representing a location in a previously returned list of files. This value allows you to start a new getmiglist request from a particular point in a previously returned list. Only files with cookies greater than or equal to cookie are included in the new list. This is an optional value. The default returns all files regardless of their cookie.
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Command reference
Table 55
Options for getmiglist (page 2 of 2)
Option
Description
-n num
The -n option takes the argument num which is the maximum number of files to include on the list. This is an optional value. The default returns all files.
-s size
The -s option takes the argument size which is the size of a file in bytes. Only files bigger than or the same as size are included on the list. This is an optional value. The default returns all files.
-k
The -k option causes getmiglist to return a summary only. The summary includes the total number of files that need to be migrated and the total combined size of the migration candidates, in bytes. This is an optional value. The default returns the file list.
List format The getmiglist file list is in the following format: Migration Weight-Blocks-Cookie-Handle-Path
The meaning of each column is described in Table 56 on page 193.
Table 56
getmiglist file list columns
Option
Description
Migration Weight
Represents how likely a file is to be migrated relative to the other migration candidates.
Blocks
Number of file system blocks used to hold the file.
Cookie
Integer that represents a file’s location on the file list.
Handle
Unique identifier for a file.
Path
Full path of the file relative to the mount point of the file system. If this cannot be determined, the list displays the string "UNKNOWN".
getpurgelist
The getpurgelist command retrieves a list of files that need to be purged. Syntax
getpurgelist -p fspath [-a age] [-c cookie] [-n num] [-s size] [-k]
193
Command reference
Options The options for getpurgelist are described in Table 57 on page 194. Table 57
Options for getpurgelist
Option
Description
-p fspath
The -p option takes the argument fspath which is the full path of the managed file system. This value is required.
-a age
The -a option takes the argument age which is the age in minutes since a file was last accessed or modified. All files accessed or modified within the time specified by age are excluded from the list. This is an optional value. The default is 0 which ignores when a file was last accessed or modified.
-c cookie
The -c option takes the argument cookie which is an integer representing a location in a previously returned list of files. This value allows you to start a new getpurgelist request from a particular point in a previously returned list. Only files with cookies greater than or equal to cookie are included in the new list. This is an optional value. The default returns all files regardless of their cookie.
-n num
The -n option takes the argument num which is the maximum number of files to include on the list. This is an optional value. The default returns all files.
-s size
The -s option takes the argument size which is the size of a file in bytes. Only files bigger than or the same as size are included on the list. This is an optional value. The default returns all files.
-k
The -k option causes getpurgelist to return a summary only. The summary includes the total number of files that need to be purged and the total combined size of the purge candidates, in bytes. This is an optional value. The default returns the file list.
List format The getpurgelist file list is in the following format: Purge Weight-Blocks-Cookie-Handle-Path
The meaning of each column Heading is described in Table 58 on page 195.
194
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Command reference
Table 58
getpurgelist file list columns headings
Heading
Description
Purge Weight
Represents how likely a file is to be purged relative to the other purge candidates.
Blocks
Number of file system blocks used to hold the file.
Cookie
Integer representing a file’s location on the file list.
Handle
Unique identifier for a file.
Path
Full path of the file relative to the mount point of the file system. If this cannot be determined the list displays the string "UNKNOWN".
migin
Use the migin command to retrieve purged file data from the back-end system. The migin command takes the argument filelist, which is the full path description of one or more files. For multiple files, separate each path description by a space. For migin examples, refer to “How to manually retrieve files” on page 165. Syntax migin filelist
migout
Use the migout command to migrate file data to the back-end system. The migout command takes the argument filelist, which is the full path description of one or more files. For multiple files, separate each path description by a space. For migout examples, refer to “How to manually migrate a file” on page 164. Syntax migout filelist
migpurgestate
Use the migpurgestate command to display the migration and purge status of a file. For each file specified, migpurgestate displays one of the following messages: ◆
file is not a file.
◆
file is not migrated and not purged.
195
Command reference
◆
file is migrated and not purged.
◆
file is migrated and purged.
where file is the full path of the specified file. The migpurgestate command takes the argument filelist, which is the full path description of one or more files. For multiple files, separate each path description with a space. Syntax migpurgestate filelist
purge
Use the purge command to remove file data from a file system. Only files which have been successfully migrated are eligible to be purged. Purging a file removes all of a file’s data except for a portion from the head of the file equal to the value of the Stub Size attribute. For more information about Stub Size, refer to “Configuring a file system” on page 57. The space occupied by the purged data is available for use by other files. The purge command takes the argument filelist, which is the full path description of one or more files. For multiple files, separate each path description with a space. For purge examples, refer to “How to manually purge files” on page 164. Syntax purge filelist
Backup and recovery commands The backup and recovery commands are described in the following sections:
196
◆
“dxcliplink” on page 197
◆
“dxuldm_tar” on page 199
◆
“dxbuildtar” on page 197
◆
“dmattrrecoverfs” on page 200
◆
“get_freeze_timeout” on page 201
◆
“getrecoverlist” on page 201
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Command reference
dxbuildtar
◆
“set_freeze_timeout” on page 202
◆
“utdm_fs_freeze” on page 203
◆
“utdm_fs_thaw” on page 204
◆
“utdm_recdmattrf” on page 204
The dxbuildar command is used to create an archive file, in tar format, that can be used to load objects into a UTDM file system. This command takes the output of dxcliplink (for a Centera module EMC Centera back-end) and creates a tar format file. The resulting file is used with dxdmload to create stub files that point to the back-end data. This command is located at: /opt/fsm/adm/smtodm/dxbuildtar
where /opt/fsm is the full path to the FSM installation directory. Syntax dxbuildar -f output_file -p prefix
Options The options for dxbuildtar are described in Table 59 on page 197. Table 59
Options for dxbuildtar
Option
Description
-f output_file
The -f option takes the argument output_file, which is the full pathname of the resulting archive file.
-p prefix
The -p option takes the argument prefix, which is the segment of the pathnames produced by dxcliplink that is not part of the path for the objects in the UTDM file system.
Example 8 on page 198 provides an example of this command. dxcliplink
The dxcliplink command parses an EMC Centera C-Clip and produces a file containing the attribute information that is needed to establish a UTDM file system link to the data represented by the C-Clip. Syntax dxcliplink -p pathname -o outputfile
197
Command reference
Options The options for dxcliplink are described in Table 60 on page 198.
Table 60
Options for dxcliplink
Option
Description
-p pathname
The -p option takes the argument pathname, which is the full path to the mountpoint of a UTDM file system that is configured with an EMC Centera back-end.
-o outputfile
The -o option takes the argument outputfile, which is the full path of the output file.
Example 8
Recovering data from an EMC Centera back-end
In this example a connection to the data represented by two C-Clip files, ABC and DEF, must be restored to a UTDM file system mounted at /mnt/utdm.1. The pathname contains the segment /mnt/utdm.1 which is the segment of the full pathnames for ABC and DEF that is not needed in the pathname relative to the UTDM file system. 1. Set the FSM environment. To set the environment, refer to “Setting the FSM environment” on page 177. 2. Create the file /tmp/dxclips.in that contains each C-Clip ID, one per line separated by a line feed (LF) character. 3. Run dxcliplink to obtain the necessary C-Clip attribute information: dxcliplink -p /mnt/utdm.1 -o /tmp/dxclipinfo.out < /tmp/dxclips.in
4. Run dxbuildtar to generate a tar-format file containing the attribute information from the C-Clips: dxbuildtar -p /mnt/utdm.1 -f file.tar < /tmp/dxclipinfo.out
5. Run dxdmload to load the attribute information into the UTDM file system: dxdmload -xUfv file.tar
dxdmload
Use the dxdmload command to load the information in a file created by dxbuildtar into a UTDM file system. This command is located at: /opt/fsm/adm/smtodm/dxdmload
198
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Command reference
where /opt/fsm is the full path to the FSM installation directory. Syntax dxdmload -xUfv dxtarfile
Options The options for dxdmload are described in Table 61 on page 199. Table 61
Options for dxdmload
Option
Description
-x
The -x option causes dxdmload to extract the information in the file specified by the -f option into the UTDM file system specified by the same file.
-U
The -U option ensures that files that exist in the UTDM file system are unlinked and recreated rather than being overwritten.
-f dxtarfile
The -f option takes the argument dxtarfile, which is a tar-formatted file created by dxbuildtar.
-v
The -v option produces verbose output.
Example 8 on page 198 provides an example of this command. dxuldm_tar
Use the dxuldm_tar command to back up and recover data in file systems. Data that exists on the back-end system is not backed up by this command. Only data resident in a file system, and not migrated, is backed up. Always exclude the .DMATTR files from backup by using the --exclude option. The dxuldm_tar command is based on the GNU version of tar. However, the command is enhanced so that it determines the file system objects required for backup. For step-by-step instructions on the use of dxuldm_tar, refer to “FSM tools” on page 143. Syntax dxuldm_tar --exclude=.DMATTR -cvf tarfile [dir|.] dxuldm_tar -xvf tarfile [filelist]
Options The options for dxuldm_tar are described in Table 62 on page 200.
199
Command reference
Table 62
Options for dxuldm_tar
Option
Description
-cvf tarfile
The -c option instructs the dxuldm_tar command to create an archive. The -v option displays filenames as they are added to the archive file. The -f option takes two arguments: The first argument is tarfile, which is the full path description for the location where the archive should be created. The second argument is dir, which is a the full path description of the top-level directory to be archived. When dir is a directory, dxuldm_tar recursively archives all subdirectories and files beneath the given directory. Optionally, a period character at the end of the command can be used to indicate the current working directory.
-xvf tarfile
The -x option instructs the command to extract an archive. The -v option displays filenames as they are extracted. The -f option takes the argument tarfile, which is the full path description for an archive to extract. The archive is extracted into the current working directory.
filelist
The dxuldm_tar command takes the argument filelist, which is the full path description of one or more files. For multiple files, separate each path description by a space. When filelist is provided, dxuldm_tar only extracts the specified files. If filelist includes a full path to a directory which matches a directory path in the archive, then the directory is recursively extracted.
dmattrrecoverfs
The dmattrrecoverfs command synchronizes the metadata for recovered migrated files with the corresponding data on the back-end system. Use this command after a file system is recovered by using NetWorker software. The dmattrrecoverfs command reestablishes the link between each recovered migrated file and the file’s data on the back-end system. For NetBackup and FSM backup tools, recovered migrated files have the recover bit set. When a file with the recover bit set is accessed, dmattrrecoverfs is automatically called to synchronize the file’s data. The dmattrrecoverfs command is also started by an hourly cronjob that is installed in root’s crontab. When dmattrrecoverfs is started it scans the file system for files with the recover bit set and synchronizes any it finds. The dmattrrecoverfs command can be run while the file system is in use.
200
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Command reference
Syntax dmattrrecoverfs -p dirpath [-i threads]
Options The options for dmattrrecoverfs are described in Table 63 on page 201. Table 63
Options for dmattrrecoverfs
Option
Description
-p dirpath
The -p option takes the argument dirpath, which is the full path to the top-level directory to be scanned. This parameter is required.
-i threads
The -i option takes the argument threads, which is the number of parallel forked threads dmattrrecoverfs can start. The range is between 1 and 64. The default is 1.
get_freeze_timeout
The get_freeze_timeout command retrieves the maximum number of seconds that a file system may be frozen by using utdm_fs_freeze. This number is called the maximum freeze timeout. The maximum freeze timeout default value is 300 seconds, but it may be changed on a host-by-host basis by using set_freeze_timeout. Syntax get_freeze_timeout
getrecoverlist
The getrecoverlist command retrieves a list of files that need to be recovered. Syntax getrecoverlist -p fspath [-k]
Options The options for getrecoverlist are described in Table 64 on page 202.
201
Command reference
Table 64
Options for getrecoverlist
Option
Description
-p fspath
The -p option takes the argument fspath which is the full path of the managed file system. This value is required.
-k
The -k option causes getrecoverlist to return a summary only. The summary includes the total number of files that need to be recovered and the total combined size of the recovery candidates, in bytes. This is an optional value. The default returns the file list.
List format The getrecoverlist file list is in the following format: Cookie-Handle-Path
The meaning of each column heading is described in Table 65 on page 202. Table 65
getrecoverlist file list columns
Heading
Description
Cookie
Integer representing a file’s location on the file list.
Handle
Unique identifier for a file.
Path
Full path of the file relative to the mount point of the file system. If this cannot be determined the list displays the string "UNKNOWN".
set_freeze_timeout
The set_freeze_timeout command sets the maximum number of seconds that a file system may be frozen by using utdm_fs_freeze. This number is called the maximum freeze timeout. The maximum freeze timeout default value is 300 seconds, the range of permissible values is 0 seconds to 600 seconds. Syntax set_freeze_timeout maxtime
Option The set_freeze_timeout command takes one required argument, maxtime, which is the new maximum freeze timeout in seconds.
202
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Command reference
utdm_fs_freeze
The utdm_fs_freeze command quiesces, suspends access to, a mounted UTDM file system, for a specified number of seconds. This allows you to generate a snapshot of a mounted file system without the possibility of file system activity causing a corrupted snapshot. At the end of the specified time period activity on the file system is automatically resumed, that is, made available for normal use. The snapshot must be completed before the file system is thawed to avoid a corrupted snapshot. You can resume file system activity before the expiration of the suspension period by using utdm_fs_thaw. Note: The utdm_fs_freeze command should only be used on mounted file systems.
Syntax utdm_fs_freeze [-t timeout] mountpoint
Options The options for utdm_fs_freeze are described in Table 66 on page 203.
Table 66
Options for utdm_fs_freeze
Option
Description
-t timeout
The -t option takes the argument timeout which is the number of seconds the file system remains frozen. At the expiration of timeout seconds, the file system is automatically thawed. This parameter is optional. The default is the maximum freeze timeout. If specified it must be less than that value. The default maximum freeze timeout is 300 seconds. For more information about the maximum freeze timeout, refer to “get_freeze_timeout” on page 201 and “set_freeze_timeout” on page 202.
mountpoint
The utdm_fs_freeze command takes the argument mount point which is the full path to the mount point for the file system. This value is required.
203
Command reference
Freeze described When you run utdm_fs_freeze, the following actions are taken for the specified file system: ◆
Current operations are allowed to finish.
◆
Any thread attempting to access the file system is suspended.
◆
Dirty buffers are flushed.
◆
A stable image of the file system is created on the disk.
Any thread that was suspended by utdm_fs_freeze remains suspended until the file system is thawed, either by the expiration of the specified time period or by running utdm_fs_thaw. utdm_fs_thaw
The utdm_fs_thaw command resumes activity on a mounted UTDM file system that was quiesced with utdm_fs_freeze. This command resumes all threads that were suspended when the file system was quiesced. Syntax utdm_fs_thaw mountpoint
Option The utdm_fs_thaw command takes one required argument, mountpoint, which is the full path to the mount point for the file system. utdm_recdmattrf
!
The utdm_recdmattrf command recovers a file system that is restored from a snapshot. This command initiates the process of creating a valid namespace for each migrated file in the recovered snapshot. IMPORTANT The utdm_recdmattrf command must be run on the file system before it is used in read/write mode. The utdm_recdmattrf command processes the restored file system’s .DMATTR file and sets the Recover attribute for each migrated file. The core processes detect the Recover attribute on a file and create a new namespace for the file during the first operation on it. The Recover attribute is then unset.
204
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Command reference
Syntax utdm_recdmattrf pathname
Option The utdm_recdmattrf command takes one required argument, pathname, which is the full path of the recovered file system’s .DMATTR file.
205
Command reference
206
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Index
Symbols
B
.deleted extension disk resource 157 .DMATTR file backup 139, 143 dxuldm_tar 199 expand_dmattrf 186 NetBackup exclude_list 141 NetBackup recovery 141 NetWorker 133 utdm_recdmattrf 205
backend Centera 64 clean-up 167 deleted files 157 FTP 67 IDM 64 module 44 mount point 54 NFS 54, 68 types 44 backup 123 .DMATTR file 133, 139, 143 commands 196 FSM tools 143 FSM-related files 124 FTP module file system 130 general 123 NetBackup 137, 141 NetWorker 132 NetWorker, 64-bit Linux version 136 recommendations 124 software 126 UTDM file systems 125 utilities 196
A actions Centera 84 disk resource 85 NFS 85 activities domain 32 file system 37 adding, user 26 administration Console 26 files 164 administrative commands 178 Agent, Console 20 associativity 78 audience 9
C Centera actions 84 backend 64 module 45 See also EMC Centera changing password, user 28, 29
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
207
Index
check directory, NFS 55 clean-up backend 167 disk resource 170 EMC Centera 168 find 170 periodic 167 storage target 170 clearing events 30 Client, Console 24 command reference administrative 178 backup 196 dx_read_log 179 dxbuildtar 197 dxcliplink 197 dxdmload 198 dxuldm_tar 197 dxuldmcenteraping 182 dxuldmcheckconf 183 dxuldmclip 183 dxuldmdelclips 184 dxuldmrecover 135 expand_dmattrf 186 file management 191 get_freeze_timeout 201 getdmattr 187 getfileattr 188 getmiglist 192 getpurgelist 193 getrecoverlist 201 migin 195 migout 195 migpurgestate 195 prtdmsession 188 purge 196 quick reference 174 set_freeze_timeout 202 utdm_fs_freeze 203 utdm_fs_thaw 204 utdm_recdmattrf 204 commands administrative 178 backup 196 internal 158 quick reference 174 recovery 196 208
usage 177 completion message, purge 165 compression levels 61 partial read size 62 configuration deleting 72 disk resource 68 editing 71 file system 43, 57 NFS 68 viewing 70 configuring file system 57 connect addresses, dxuldmdelclips 185 connectivity EMC Centera 46 FTP 52 Console administration 26 overview 21 starting Client 24 console processes 162 Console Agent 20 Console Client 24 IPv6 24, 25 monitoring 30 starting 24 console processes logs 162 monitoring 162 stopping 162 console processes , starting 162 Console Server 20 conventions 10 copy, extended rules 89 core processes general 156 monitoring 160 create, extended rules 80
D data management utilities 191 data retention EMC Centera 51, 111 EMC Centera requirements 111
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Index
general 91, 109 deldmsession 178 deleted files backend 157 general 157 migdestroy 167 deleting configuration 72 extended rules 88 determining maximum time-out 148 direct-read 120 memory-mapping 120 mounting 121 disk resource actions 85 clean-up 170 configuration 68 preliminary tasks 54 dmap_root_path, setting 177 DMAPI 14 DMAPI attributes file, snapshot 204 dmattrrecoverfs 200 domain activities 32 events 31 dx_read_log command reference 179 monitoring processes 161 options 181 severity levels 181 dxbuildtar 197 dxcliplink 197 dxdmload 198 dxuldm_tar command reference 197 options 200 using 143 dxuldmcenteraping, command reference 182 dxuldmcheckconf, command reference 183 dxuldmclip 183 dxuldmdelclips command reference 184 connect addresses 185 EMC Centera 168 purge 184 restart 169
dxuldmrecover 135
E editing configuration 71 extended rules 87 effect of manually stopping, processes 159 EMC Centera clean-up 168 connectivity 46 data retention 51, 111 data retention requirements 111 dxcliplink 197 dxuldmdelclips 168 IPv4 64 PEA file 47 PEA file permissions 168, 185 permissions 46 preliminary tasks 45 purge 167 EMC Centera data retention enabling 111 performance 110 setting for a file system 113 setting for files 114 EMC Centera, IPv4 46, 48 enable data compression 61, 62 events clearing 30 domain 31 file system 35 host 34 example migin 166 migout 164 expand_dmattrf, command reference 186 exporting, extended rules 89 expressions 74 extended rules 73 copy 89 copying 89 create 80 creating 74 deleting 88 editing 87 exporting 89
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
209
Index
viewing 86
F file deletion, uvdmigd 167 file management, command reference 191 file system activities 37 configuration 43, 57 configuring 57 events 35 recovery 128 retention 113 state 38 files administration 164 deleted 157 management commands 191 manually migrate 164 manually purge 164 manually retrieve 165 retention 114 find, clean-up 170 freeze 147 maximum time-out 148 snapshot 147 SnapView 147 UTDM file systems 147 FSM internal commands 158 migpurged 156 monitoring 160 starting 158 stopping 159 uvdmigd 157 FSM environment, setting 177 FSM tools backup 143 recover 145 FSM-aware software 127 FSM-related files, backup 124 FTP backend 67 connectivity 52 IPv6 52, 67 preliminary tasks 52 username and password 52
210
G get_freeze_timeout command reference 201 snapshot 201 SnapView 201 getdmattr, command reference 187 getfileattr, command reference 188 getmiglist, command reference 192 getpurgelist, command reference 193 getrecoverlist, command reference 201
H host, events 34 hostname database 45 hosts 45
I IDM backend 64 preliminary tasks 45 incremental backups, recovery 115 individual filesystems, monitoring 164 internal commands, FSM 158 internal, commands 158 IPv4 24, 25, 46, 48, 64, 67 IPv6 52, 67 Console Client 24, 25
L logs console processes 162 core processes 161
M maintenance 155 management commands, files 191 management domain 22 management utilities 191 manual migrate 164 purge 164 retrieve 165 starting FSM processes 158 stopping FSM processes 159
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Index
manual data retrieval, uvdmigd 165 manually migrate, files 164 manually purge, files 164 manually retrieve, files 165 manually starting, processes 158 manually stopping, processes 159 match string 74 maximum time-out determining 148 freeze 148 setting 148 snapshot 148 migdestroy deleted files 167 process 157 migin command reference 195 example 166 uvdmigd 157 migout command reference 195 example 164 migpurged FSM 156 process information 156 migpurgestate, command reference 195 migrate, manual 164 minimum topology 17 module backend 44 Centera 45 monitoring Console Client 30 console processes 162 core processes 160 FSM 160 individual filesystems 164 monitoring processes, dx_read_log 161 mount options 109 direct-read 120 memory-mapping 120 mounting 121 EMC Centera data retention enabling 111 performance 110 setting for a file system 113 setting for files 114
read-only 117 mounting 118 readthru 120 utdmro 117 mount point, backend 54
N NetBackup .DMATTR file 141 backup 137, 141 exclude_list 141 recovery 137, 142 requirements 139 user-initiated backup 141 user-initiated restore 141 NetWorker .DMATTR file 133 backup 132 path information 132 permissions 132 recovery 133 required file 132 requirements 132 NFS actions 85 check directory 55 configuration 68 preliminary tasks 54 NFS module 54, 68
O online help 25 options dx_read_log 181 dxuldm_tar 200 overview, Console 21
P parentheses 74 path information, NetWorker 132 PEA file 47 performance 110 periodic cleanup 167 periodic, clean-up 167 permissions
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
211
Index
EMC Centera 46 EMC Centera PEA file 168, 185 NetWorker 132 precedence 78 preliminary tasks 44 AVALOMidm 45 disk resource 54 EMC Centera 45 FTP 52 IDM 45 NFS 54 process information migpurged 156 uvdmigd 157 processes 156 console 162 effect of manually stopping 159 manually starting 158 manually stopping 159 uvdmigd 157 prtdmsession, command reference 188 prune job 95 purge command reference 196 completion message 165 dxuldmdelclips 184 EMC Centera 167 manual 164
Q quick reference commands 174 utilities 174 quick reference, utilities 174 quiesce. See freeze
R read-only 117 mounting 118 recommendations, backup 124 recover, FSM tools 145 recovery 123 commands 196 file system 128 FTP module file system 130 general 123 212
incremental backups 115 native FTP 130 NetBackup 137, 142 NetWorker 133 NetWorker, 64-bit Linux version 136 single file 145 snapshot 149 software 126 removing, user 28 required file, NetWorker 132 requirements NetBackup 139 NetWorker 132 restart, dxuldmdelclips 169 restoring incremental backups 115 retention file system 113 files 114 retrieve, manual 165
S scheduled jobs comparison 92 delete old files 92, 97 delete old files job, backend differences 94 deleting 100 editing 99 effect on existing settings 93 job types 92 migrate 92 overview 92 prune 92, 95, 97 purge 92 reoccurrence options 102 daily 104 monthly 105 one specific time 104 weekdays 106 weekends 105 weekly 105 yearly 106 schedule components 93 time element creating 102 deleting 107 editing 107
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
Index
overview 102 viewing 98 Server, Console 20 set_freeze_timeout command reference 202 snapshot 202 SnapView 202 setting dmap_root_path 177 FSM environment 177 maximum time-out 148 stub size 166 severity levels, dx_read_log 181 single file, recovery 145 snapshot 127 DMAPI attributes file 204 freeze 147 get_freeze_timeout 201 maximum time-out 148 recovery 149 set_freeze_timeout 202 SnapView 147 thaw 149 utdm_fs_freeze 203 utdm_fs_thaw 204 utdm_recdmattrf 204 SnapView freeze 147 get_freeze_timeout 201 set_freeze_timeout 202 snapshot 147 thaw 149 utdm_fs_freeze 203 utdm_fs_thaw 204 utdm_recdmattrf 204 software backup 126 recovery 126 Solaris zones 54 back-end mount point 55 starting Console Client 24 console processes 162 FSM 158 starting Client, Console 24 starting FSM processes, manual 158
state, file system 38 stopping console processes 162 FSM 159 stopping FSM processes, manual 159 storage target, clean-up 170 stub size setting 166
T thaw snapshot 149 SnapView 149 utdm file system 149 time elements 102 creating 102 topology 17 topology guidelines 18 types, backend 44
U usage, commands 177 user adding 26 changing password 28, 29 removing 28 user-initiated backup, NetBackup 141 user-initiated restore, NetBackup 141 username and password, FTP 52 utdm file system, thaw 149 UTDM file systems backup 125 freeze 147 utdm_fs_freeze command reference 203 snapshot 203 SnapView 203 utdm_fs_thaw command reference 204 snapshot 204 SnapView 204 utdm_recdmattrf .DMATTR file 205 command reference 204 snapshot 204 SnapView 204
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
213
Index
utilities administrative 178 backup 196 backup and recovery 196 file management 191 quick reference 174 uvdmigd file deletion 167 FSM 157 manual data retrieval 165 migin 157 process information 157 processes 157
V viewing configuration 70 extended rules 86 logs 161
Z zones, Solaris 54
214
EMC DiskXtender File System Manager for UNIX/Linux Release 3.5 Administrator’s Guide
