Transcript
Technical white paper
HP MSA 2040 Storage 750 Mailbox Resiliency Exchange 2013 Storage Solution with Microsoft Hyper-V Virtualized Microsoft Exchange 2013 server deployment
Table of contents Overview .......................................................................................................................................................................................... 2 Disclaimer........................................................................................................................................................................................ 2 Features of the tested solution .................................................................................................................................................. 2 HP MSA 2040 Storage .............................................................................................................................................................. 2 HP ProLiant DL380p Gen8 Server ......................................................................................................................................... 3 HP StoreFabric SN1000E 16Gb Host Bust Adapter ........................................................................................................... 4 Solution description ...................................................................................................................................................................... 5 Targeted customer profile ...................................................................................................................................................... 5 Storage configuration .............................................................................................................................................................. 6 Mailbox configuration – primary site operational .............................................................................................................. 7 Mailbox configuration – primary site unavailable (planned or unplanned) .................................................................. 8 Mailbox configuration – primary site offline and single server failure at Disaster Recovery site (planned or unplanned) ........................................................................................................................................................... 9 Tested deployment ................................................................................................................................................................ 10 Best practices ............................................................................................................................................................................... 11 Test results summary ................................................................................................................................................................ 12 Reliability................................................................................................................................................................................... 13 Storage performance results ............................................................................................................................................... 13 Database backup / recovery performance ........................................................................................................................ 13 Conclusion ..................................................................................................................................................................................... 14 Appendix A – stress test ............................................................................................................................................................ 15 Appendix B – performance test ............................................................................................................................................... 17 Appendix C – streaming backup test....................................................................................................................................... 19 Appendix D – soft recovery test ............................................................................................................................................... 20 For more information ................................................................................................................................................................. 22
Technical white paper | HP MSA 2040 Storage 750 Mailbox Exchange 2013 Mailbox Resiliency Storage Solution
Overview This document provides information on a 750-mailbox, Microsoft® Exchange Server 2013 (Exchange) storage solution tested on HP MSA 2040 Storage and virtualized with Microsoft Hyper-V on HP ProLiant DL380p Gen8 servers according to criteria specified by Microsoft’s Exchange Solution Reviewed Program (ESRP) – Storage program. For any questions or comments regarding the contents of this document, see the For more information section. The ESRP – Storage program is a Microsoft Exchange Server program designed to facilitate third -party storage testing and solution publishing for Exchange Server. For more details on the program, see: technet.microsoft.com/en-us/exchange/ff182054.aspx Tested with: ESRP – Storage Version 4.0 Test date: November 2013
Disclaimer HP produced this document independently of Microsoft Corporation. Microsoft Corporation expressly disclaims responsibility for, and makes no warranty, express or implied, with respect to the accuracy of the contents of this document.
Features of the tested solution This solution is for small and midsize businesses that are planning to deploy and virtualize Exchange on HP ProLiant DL380p Gen8 Servers and HP MSA 2040 Storage. The solution allows the use of the Exchange mailbox resiliency database availability group (DAG) feature with the high-availability features of the HP MSA 2040 Storage; providing hardware efficiency and high availability for Exchange users. The solution design is a virtualized Exchange environment supporting 750 mailboxes in a resiliency configuration across two sites (primary site and disaster recovery site). The solution uses a single DAG to support 750 mailboxes, each with a 5 GB mailbox capacity and a work profile of 200 messages sent and received for each user, each day, with additional performance and capacity overhead included. To provide mailbox resiliency, there are four copies of each database; two copies located at the primary site and two copies located at the disaster recovery site. The DAG contains four mailbox servers (two at the primary site and two at the disaster recovery site), with each mailbox server configured as a virtual machine. Each virtual machine resides on its own hypervisor, using the virtualization features of Microsoft Hyper-V. The solution eliminates a single point of failure situation and handles the loss of a storage array, hypervisor server, mailbox server virtual machine, database volume, host bus adaptor (HBA), or switch. For example, during a mailbox server failure or maintenance operation, each mailbox server can handle the compute and storage requirements of all 750 mailboxes. The primary site hosts all 750 mailboxes during normal operations. The performance results outlined in this document are from the testing of a single building block, with a single server failed at the disaster recovery site. See the section Test results summary for more information.
HP MSA 2040 Storage The HP MSA 2040 shown in Figure 1 is a high-performance storage array, designed for entry-level HP customers desiring 8 and/or 16Gb Fibre Channel, 1 and/or 10GbE iSCSI, or 6Gb/12Gb SAS connectivity with 4 host ports per controller. This next generation HP MSA 2040 provides excellent value for customers needing performance balanced with price to support initiatives such as consolidation and virtualization. Figure 1. HP MSA 2040 Storage shown in a large form factor (LFF) configuration
2
Technical white paper | HP MSA 2040 Storage 750 Mailbox Exchange 2013 Mailbox Resiliency Storage Solution
Key features and benefits Easily managed, highly affordable, highly efficient, and flexible storage consolidation • Provides choice of controllers to match your infrastructure configuration needs. – MSA 2040 SAN Controller supports 8Gb/16Gb FC and/or 1GbE/10GbE iSCSI – MSA 2040 SAS Controller supports 6Gb/12Gb SAS • Offers single controller models for low initial investment, or dual controllers for higher data protection. The FC controller
allows sharing of resources between different departments with the most cost-effective protocol. A wide choice of drive types means a high return on investment (ROI). • Offers a choice of drives: high-performance and high-speed SSD drives, enterprise-class SAS, and SAS Midline drives. The
array head comes with either 12 (LFF) or 24 (SFF) drive bays, able to accommodate all three drive types. • Ships standard with 64 snapshot and Volume Copy cloning capability enabled, optionally increase snapshots to 512
snaps. • All HP MSA 2040 Storage controller models support the Remote Snap (replication) software option.
Start small and grow in number of supported servers and total capacity – adaptable to changing needs • Grow storage as demands increase up to a raw capacity of 384TB with 4TB LFF drives. The HP MSA 2040 Array can support 8 LFF enclosures and 96 drives, or 8 SFF enclosures and 199 drives. Single or dual controllers accommodate any budget. • MSA 2040 SAN controller supports a combination of host interface protocols by mixing Fibre Channel and iSCSI protocols
on the same controller, with 4GB Cache achieving better application response time and the ability to support more virtualized environments. • With 512 LUNs and max volume size of 64TB, the HP MSA 2040 Array has more than sufficient LUN flexibility for a wide
variety of configuration choices. Comprehensive data protection in an easily administered SAN environment • Reduces risk of IT failure with built in dual active/active controllers, dual-ported drives, and redundant hardware components. • Increases data protection with HP MSA 2040 Array standard 64 Snapshot (optional 512) and Volume Copy software.
Allows space-efficient, copy-on-write snapshots and clones. Optional Remote Snap (replication) is available with the Fibre Channel model. • Reduces deployment time with easy and secure configuration. This reduces operator errors, resulting in less downtime
and data loss, and increased customer satisfaction. • Supports data-in-place upgrades, protecting the user’s investment in JBODs and drives, enhancing performance and
functionality. For more information on the HP MSA 2040 Storage, visit hp.com/go/MSA2040.
HP ProLiant DL380p Gen8 Server With innovation in performance, feature flexibility, and management capabilities, the HP ProLiant DL380p Gen8 Server shown in Figure 2 is an exceptional data center investment for managing today’s demanding computing requirements. Figure 2. HP ProLiant DL380p Gen8 Server
3
Technical white paper | HP MSA 2040 Storage 750 Mailbox Exchange 2013 Mailbox Resiliency Storage Solution
Key features and benefits Redefining the customer experience with HP Proactive Insight architecture • Prevents data loss, reduces downtime, and improves serviceability. Includes HP SmartDrives, HP Smart Socket guides, “Snap and Go” rail kit options, and tool-less access to components. • Provides intuitive system management with simplified provisioning, proactive health monitoring and alerting, and
automated firmware and system software maintenance, using Integrated Lifecycle Automation. • Dynamic Workload Acceleration converges storage, compute, and I/O to turbo-charge performance and resiliency while
eliminating bottlenecks. • Reclaims space, power, and cooling resources needed for workloads and extends your data center capacity with
Automated Energy Optimization. • Enables industry leading performance, uptime, and productivity, integrated into a personalized, simplified support
experience with HP support services. Unlock your infrastructure’s potential with HP’s easy-to-use integrated management and support • Offers a complete set of embedded features, standard on all ProLiant Gen8 servers, including HP iLO, HP Agentless Management, HP Active Health System, HP Intelligent Provisioning, and HP Embedded Remote Support using the HP iLO Management Engine. • Provides 24x7 remote monitoring and anywhere, anytime, personalized access to your IT and support status with HP
Insight Online and HP Insight Remote Support. • Reduces deployment time and update complexity by systematically updating server infrastructure in the data center. It
includes HP Smart Update Manager (HP SUM), Service Pack for ProLiant (SPP), and other products using HP SmartUpdate. • Provides advanced management capabilities across the lifecycle of your HP ProLiant server infrastructure with HP Insight
Control. HP iLO Advanced, a key component, provides enhanced remote server functions that reduce IT travel costs and decrease resolution time. Improved capacity and performance on compute and storage • Prevents data loss and downtime with enhanced error handling, while improving workload performance and power efficiency with HP SmartMemory. • Improves serviceability and prevents data loss with features such as icon-based status display and Do Not Remove LED
with HP SmartDrive technology. • Offers maximum performance, memory, I/O connectivity, and bandwidth for demanding workloads with The Intel® Xeon®
processor E5-2600 product family. Energy efficiency by design • Includes a 3D array of temperature sensors that help precisely control server fans to direct cooling and reduce unnecessary fan power, saving on cooling costs. • High efficiency 80+ certified HP Common Slot Power Supplies provide up to 94% efficiency (Platinum) and support HP’s
Power Discovery Services. • ENERGY STAR® qualified server configurations illustrate HP’s continued commitment to helping customers conserve
energy and save money.
HP StoreFabric SN1000E 16Gb Host Bust Adapter The HP StoreFabric SN1000E 16Gb Host Bus Adapter shown in Figure 3 is designed to support HP ProLiant Servers with PCIExpress I/O slots to connect to HP Storage Arrays using the 16/8/4 Gb Fibre Channel protocol. The HP SN1000 16Gb Host Bus Adapters bring data center infrastructure components to a higher level of performance and efficiency with the ability to deliver twice the I/O performance of 8Gb HBAs. A 16Gb FC HBA purchased today is backward compatible with 8 and 4Gb storage networks and will protect future investments. When using storage intensive applications, such as backup/restore, database transactions, virtualization, and rich media, the increased performance of the 16Gb infrastructure enables more rapid storage and retrieval of critical information. Designed for environments with greater virtual machine density and bandwidth requirements, the SN1000 16Gb HBAs enable more applications and VMs to run on a single server and Fibre Channel port, resulting in reduced cabling and a higher return on IT investment.
4
Technical white paper | HP MSA 2040 Storage 750 Mailbox Exchange 2013 Mailbox Resiliency Storage Solution
Figure 3. HP StoreFabric SN1000E 16Gb Host Bus Adapter
Key features and benefits • Delivers up to six times the IOPS and twice the data throughput (MB/s). Cuts latency by half compared to 8Gb FC HBAs. • Provides the infrastructure required for more powerful servers and data-intensive applications of the future by moving to
16Gb technology. • Enables faster storage and data retrieval with improved I/O performance when using storage-intensive applications like
backup/restore and database transactions. • Provides an HP-branded HBA solution that has undergone extensive HP interoperability testing for connecting HP
ProLiant servers into storage and networking environments. • Enables more applications to run on a single server and Fibre Channel port, resulting in reduced cabling and a higher
return on IT investment.
Solution description This document provides best practices to help administrators implement an Exchange deployment with the following requirements.
Targeted customer profile • 750 mailbox users (375 mailbox users per server) • 5 GB mailbox size • Exchange 2013 profile of 0.16 IOPS workload per user (200 messages sent and received for each user, each day, with
additional performance and capacity overhead included) • Single DAG deployment with two mailbox servers at the primary site and two mailbox servers at the disaster recovery
site, for a total of four mailbox servers across the entire solution • Four copies of each mailbox database (one active / three passive) • Background Database Maintenance (BDM) enabled • Designed for server and site resiliency • Eight 2TB 6G SAS 7.2K LFF (3.5-inch) Dual Port MDL Hard drives per array for the Exchange databases and logs • Storage deployed as RAID 1. See the Storage configuration section for more information.
5
Technical white paper | HP MSA 2040 Storage 750 Mailbox Exchange 2013 Mailbox Resiliency Storage Solution
Note Microsoft Exchange Server 2013 is a complex application that requires thorough knowledge of deployment strategies, interdependencies, and server and storage performance goals. HP has developed the HP Sizer for Microsoft Exchange to assist customers with proper server and storage sizing and configuration for their Exchange deployments. The algorithms developed and implemented in this tool are based on HP’s extensive knowledge of both Exchange and the hardware required to run it: CPU, memory, and storage subsystems. This tool is free to download and use, and is designed to allow very quick “what if” modeling of various configuration and design choices. For more information, visit hp.com/solutions/microsoft/exchange2013/sizer. In addition to the HP Sizer, HP recommends that customers use Microsoft’s Exchange 2013 Server Role Requirements Calculator. For more information, visit gallery.technet.microsoft.com/office/Exchange-2013-Server-Role-f8a61780.
Storage configuration The recommended best practice for site and server resiliency is two HP ProLiant DL380p Gen8 servers and one HP MSA 2040 Array, with eight 2TB 6G SAS 7.2K LFF (3.5-inch) Dual Port MDL hard drives at each site. For storage performance and redundancy, connect the HP ProLiant DL380p Gen8 servers using two HP StoreFabric SN1000E HBAs in each server to both controllers on the HP MSA 2040. Each HP MSA 2040 uses eight hard drives for the Exchange databases and transaction logs. For additional redundancy against hard drive failures, additional hard drives may be added to the storage and designated as global spares, further protecting against disk failures. The Exchange restore volumes are located on the servers. Figure 4 illustrates the hardware used in this solution at the primary site. Figure 4. View of the Exchange solution at primary site (duplicated at disaster recovery site)
5
2
6
3
7
NIC
PS1
POWER CAP
PS2
1
4
3
2
4
3
4
1
4
2
5
3
6
PS1
AMP STATUS
PROCS
6
5
8
7
10 12
9
1
FANS 7.2K SAS 1.0 TB
2
OVER TEMP
UID
DIMMS 2
1
1
PORT 0
1
7.2K SAS 1.0 TB
QR559A
15K SAS 300 GB
15K SAS 300 GB
PORT 0
Rear View QR559A
Front View
1
2
11
1
2
3
4
3
6
5
8
7
4
10 12
9
PS2
11
5
6
PLC B
1200W 94%
8
ProLiant DL380p Gen8
4
UID
6
3
7
NIC
PS1
POWER CAP
PS2
2
1
1
2
3
OVER TEMP
4
AMP STATUS
UID
DIMMS
1
4
3
5
FANS
2
PROCS
6
8
7
10 12
9
1
1
2
11
1
2
3
4
3
6
5
4
8
7
5
10 12
9
11
6
3
PORT 0
5
2
QR559A
1
7.2K SAS 1.0 TB
PORT 0
15K SAS 300 GB
QR559A
iLO
15K SAS 300 GB
1
4
2
5
3
6
PS1
PS2
PLC B
1200W
7.2K SAS 1.0 TB
4
8
ProLiant DL380p Gen8
94% 4
UID
SAS MDL
6GB DP 7.2K
2.0 TB
SAS MDL
2.0 TB
SAS MDL
2.0 TB
SAS MDL
6GB DP 7.2K
2.0 TB
SAS MDL
2.0 TB
SAS MDL
2.0 TB
6Gb/s
PORT 1
PORT 2
PORT 3
6GB DP 7.2K
6GB DP 7.2K
CLI
PORT 4
SERVICE-2
CLI
ACT
LINK
SERVICE-1 LINK
UID
6GB DP 7.2K
SAS MDL
6GB DP 7.2K
2.0 TB
SAS MDL
2.0 TB
SAS MDL
2.0 TB
SAS MDL
6GB DP 7.2K
2.0 TB
SAS MDL
2.0 TB
SAS MDL
6GB DP 7.2K
6Gb/s
2.0 TB
1
4
7 10
3
6
9 12
PORT 1
PORT 2
PORT 3
PORT 4 CACHE
CLI
SERVICE-2
CLI
ACT
LINK
SERVICE-1 LINK
HP MSA 2040
6Gb/s
6Gb/s
SERVICE IN
OUT
6GB DP 7.2K
6GB DP 7.2K
LINK
LINK
UID
6GB DP 7.2K
6Gb/s 6GB DP 7.2K
SAS MDL
6GB DP 7.2K
2.0 TB
SAS MDL
6GB DP 7.2K
2.0 TB
SAS MDL
2.0 TB
6GB DP 7.2K
SAS MDL
6GB DP 7.2K
SAS MDL
6GB DP 7.2K
6GB DP 7.2K
2.0 TB
SAS MDL
2.0 TB
2.0 TB
SAS MDL
6GB DP 7.2K
2.0 TB
SAS MDL
6GB DP 7.2K
2.0 TB
SAS MDL
2.0 TB
6GB DP 7.2K
SAS MDL
6GB DP 7.2K
2.0 TB
SAS MDL
6GB DP 7.2K
2.0 TB
SAS MDL
6GB DP 7.2K
HP MSA 2040
CACHE
2
2.0 TB
TOP
1 iLO
2
TOP
1
1
4
7 10
3
6
9 12
6Gb/s
SERVICE IN LINK
OUT LINK
4 1. 2. 3. 4.
6
Two HP ProLiant DL380p Gen8 Servers Two HP MSA 2040 Storage arrays, each with eight 2TB SAS LFF drives dedicated to the Exchange databases and logs. The remaining drives are either global spares or unused. Primary and secondary paths (16Gb FC) Primary and secondary paths (12Gb SAS)
3
Technical white paper | HP MSA 2040 Storage 750 Mailbox Exchange 2013 Mailbox Resiliency Storage Solution
Mailbox configuration – primary site operational Figure 5 shows the mailbox layout for active and passive database copies when the primary site is operational. This solution brief provides best practices to help administrators implement the Exchange deployment, with the following requirements: • 750 mailbox users. • 0.16 IOPS workload per user (200 messages sent and received per user, per day, with additional performance and
capacity overhead included). • 5 GB mailbox size. • Each mailbox server in the primary site hosts four database copies (two active and two passive). • Each mailbox server in the disaster recovery site hosts four database copies (four passive). • Background database maintenance (BDM) enabled. Figure 5. Mailbox database layout during normal operations
Primary Site Virtualized Exchange Server
MBX1
Disaster Recovery Site
MBX2
MBX3
MBX4
DAG1 Database Availability Group
Active Database DB1
DB1
DB1
DB1
DB2
DB2
DB2
DB2
DB3
DB3
DB3
DB3
DB4
DB4
DB4
DB4
Passive Database
7
Technical white paper | HP MSA 2040 Storage 750 Mailbox Exchange 2013 Mailbox Resiliency Storage Solution
Mailbox configuration – primary site unavailable (planned or unplanned) Figure 6 shows the mailbox layout for active and passive database copies when the primary site is unavailable due to a planned or unplanned event. This solution brief provides best practices to help administrators implement the Exchange deployment, with the following requirements: • 750 mailbox users. • 0.16 IOPS workload per user (200 messages sent and received per user, per day, with additional performance and
capacity overhead included) • 5 GB mailbox size. • Primary site’s database copies are unavailable due to a planned or unplanned event. • Each mailbox server in the disaster recovery site hosts four database copies (two active and two passive). • Background database maintenance (BDM) enabled. Figure 6. Mailbox database layout with primary site unavailable/offline
Primary Site Virtualized Exchange Server
Database Availability Group
Active Database
MBX1
Disaster Recovery Site
MBX2
XX
MBX3
MBX4
DAG1
DB1
DB1
DB1
DB1
DB2
DB2
DB2
DB2
DB3
DB3
DB3
DB3
DB4
DB4
DB4
DB4
Passive Database
8
Technical white paper | HP MSA 2040 Storage 750 Mailbox Exchange 2013 Mailbox Resiliency Storage Solution
Mailbox configuration – primary site offline and single server failure at Disaster Recovery site (planned or unplanned) Figure 7 shows the mailbox layout for active database copies when the primary site and one of the servers in the disaster recovery site are unavailable due to a planned or unplanned event. This solution demonstrates the worst-case scenario with one host in the disaster recovery site hosting four active databases, with the following requirements: • 750 mailbox users. • 0.16 IOPS workload per user (200 messages sent and received per user, per day, with additional performance and
capacity overhead included). • 5 GB mailbox size. • Primary site’s database copies and one of the two servers at the disaster recovery site are unavailable due to a planned
or unplanned event. • The mailbox server in the disaster recovery site hosts four database copies (four active). • Background database maintenance (BDM) enabled.
The performance results in this document are based on this worst-case scenario shown in Figure 7. Figure 7. Mailbox database layout with primary site unavailable/offline and a single server failure at the disaster recovery site
Primary Site Virtualized Exchange Server
MBX1
Disaster Recovery Site
MBX2
MBX3
MBX4
XX X DAG1
Database Availability Group
Active Database
DB1
DB1
DB1
DB1
DB2
DB2
DB2
DB2
DB3
DB3
DB3
DB3
DB4
DB4
DB4
DB4
Passive Database
9
Technical white paper | HP MSA 2040 Storage 750 Mailbox Exchange 2013 Mailbox Resiliency Storage Solution
Tested deployment The following tables summarize the tested environment. Simulated Exchange configuration Number of Exchange mailboxes simulated
750 (5 GB mailbox size)
Number of Database Availability Groups (DAGs)
1
Number of servers per DAG
4
Number of databases per server
4
Number of copies per database
4
Number of mailboxes per database
188
Simulated profile: I/O operations per second per mailbox (IOPS include 20% headroom)
0.16
Database and log LUN size
1 TB
Total database size for performance testing
4 TB
Percent of storage capacity used by Exchange databases
55 % 4 TB/7.24 TB (Total database size /Total formatted capacity)
Storage hardware Storage connectivity
Fibre Channel (FC)
Storage model and firmware version
HP MSA 2040 SAN Storage Firmware Version: GL101R001 hp.com/go/msa2040
Storage cache
4 GB per storage controller
Number of storage controllers
2
Number of storage ports
4 x 16/8 Gb FC ports per storage controller
Maximum bandwidth of storage connectivity to tested configuration
2 x 16Gb/s
Host HBA model and firmware version
HP StoreFabric SN1000E 16Gb HBA Firmware Version: 1.1.43.6
Number of HBAs per server
2
Server type
HP ProLiant DL380p Gen8 Server CPU: (2) Intel Xeon E5-2690 @ 2.90GHz Memory: 192 GB
10
Total number of disks tested in solution
8
Maximum number of disks that can be hosted in the tested configuration
12
Technical white paper | HP MSA 2040 Storage 750 Mailbox Exchange 2013 Mailbox Resiliency Storage Solution
Storage software HBA driver and driver version
Emulex Storport Miniport Driver Driver Version: 2.76.002.001
HBA QueueTarget setting
0 (default)
HBA QueueDepth setting
32 (default)
Multipathing
Microsoft DSM (MPIO) MPIO Policy: round robin with subset
Host OS and hypervisor provider
Microsoft Windows® Server 2012 Datacenter Hypervisor: Microsoft Hyper-V
Guest OS
Microsoft Windows Server 2012 Datacenter vCPU: 6 Memory: 48 GB
Jetstress version
15.00.0775.000
ESE version
15.00.0626.000
Storage disk configuration (mailbox and log disks) Disk type, speed, and firmware version
2TB 6G SAS 7.2K LFF (3.5-inch) Dual Port MDL Hard Drive Firmware Version: A001
Raw capacity for each disk
2 TB
Number of physical disks in test
8
Total raw storage capacity
16 TB
RAID level
RAID 1
Total formatted capacity
7.24 TB
Storage capacity utilization
45 % 7.24 TB / 16 TB (Formatted capacity/ Total Raw capacity)
Database capacity utilization
25 % 4 TB / 16 TB (Total database size / Total Raw capacity)
Replication configuration Replication mechanism
Mailbox resiliency
Number of links per guest OS
2
Simulated link distance
LAN
Link type
IP
Link bandwidth
1 Gb
Best practices The tested solution used the following best practices and recommendations. • Drivers and Firmware
11
Technical white paper | HP MSA 2040 Storage 750 Mailbox Exchange 2013 Mailbox Resiliency Storage Solution
Ensure that firmware and drivers for all hardware in the solution are up to date. For more information, see: hp.com/go/support. • Hyper-Threading
Turn it off. Hyper-Threading is known to improve CPU throughput for most applications. However, the benefits to Exchange 2013 do not outweigh the negative impacts. It turns out that there can be a significant impact to memory usage on Exchange servers when Hyper-Threading is enabled, due to the way the .NET server garbage collector allocates heaps. • Multipathing
Install Microsoft’s Multipath I/O (MPIO). Multipathing allows for redundant physical paths to the storage and load balancing of the I/O. For more information, see: microsoft.com/en-us/download/details.aspx?id=30450. • Jetstress 2013
Before installing Exchange Server 2013, test your storage solution with Jetstress to ensure it is configured properly and meets your requirements. For more information on Jetstress, see: microsoft.com/en-us/download/details.aspx?id=36849. • RAID Level
RAID 1 is the selected RAID configuration for this solution with a stripe size of 256 KB (256 KB or larger is recommended). Provisioning of other RAID types is supported but not recommended for this Exchange Server solution. • Mailbox Database Copies
In addition to the features of RAID 1 protection, additional fault tolerance was achieved with four copies of the mailbox databases. • Hypervisor Snapshots
Some hypervisors include features for taking snapshots of virtual machines. Virtual machine snapshots capture the state of a virtual machine while it is running. This feature enables you to take multiple snapshots of a virtual machine and then revert the virtual machine to any of the previous states by applying a snapshot to the virtual machine. However, virtual machine snapshots are not application aware; using them can have unintended and unexpected consequences for a server application that maintains state data, such as Exchange. As a result, making virtual machine snapshots of an Exchange guest virtual machine is not supported. • Oversubscription
Many hardware virtualization products allow you to specify the number of virtual processors that should be allocated to each guest virtual machine. The virtual processors located in the guest virtual machine share a fixed number of logical processors in the physical system. Exchange supports a virtual processor-to-logical processor ratio no greater than 2:1, although we recommend a ratio of 1:1. For example, a dual processor system using quad core processors contains a total of 8 logical processors in the host system. On a system with this configuration, do not allocate more than 16 virtual processors to all guest virtual machines combined. • Follow Hypervisor Vendor Guidance
Always consult the Hypervisor vendor for best practices and recommendations when virtualizing Exchange deployments. For specific information on Exchange 2013 Virtualization, visit: technet.microsoft.com/en-us/library/jj619301(v=exchg.150).aspx • Windows Volumes
Ensure that the Windows volumes are formatted with the 64KB allocation unit size for best performance. For more information on Exchange 2013 Storage Configuration Options and Best Practice recommendations, see: technet.microsoft.com/en-us/library/ee832792.aspx
Test results summary This section provides a high-level summary of the test data from ESRP and a reference to the detailed HTML reports, generated by the ESRP testing framework. For solutions that use a uniform building block style, in which each building block is independent and isolated, ESRP requires testing only one building block. For this solution, a building block consisted of one HP ProLiant DL380p Gen8 server and one 12
Technical white paper | HP MSA 2040 Storage 750 Mailbox Exchange 2013 Mailbox Resiliency Storage Solution
HP MSA 2040 array, with eight (8) 2TB 6G SAS 7.2K LFF (3.5-inch) Dual Port MDL drives hosting all 750 mailboxes. However, for actual deployments of this solution, additional hardware is required at both the primary and disaster recovery sites. For a complete solution, contact your HP Reseller or HP Sales Representative.
Reliability Reliability testing exercises the storage with the maximum sustainable Exchange I/O load for 24 hours. The goal is to verify that the storage can handle the maximum I/O load for an extended amount of time. Both log and database files are analyzed for integrity after the stress test to confirm that no database or log corruption occurred. After running the 24-hour test, we found no relevant errors reported in the event log for the storage reliability testing, and no errors reported for the checksum process.
Storage performance results Storage performance testing exercises the storage with the maximum sustainable Exchange I/O load for two hours. The test measures storage I/O response times under the maximum sustainable Exchange I/O load. The following data is the sum of all logical disk I/Os and an average of all logical disk I/O latency in the two-hour test duration. This solution is designed around an Exchange deployment with a single DAG. The DAG has two mailbox servers in the primary site and two mailbox servers in the disaster recovery site. Ideally, the two servers in the disaster recovery site should host all active mailboxes when the primary site is offline (planned or unplanned). The storage performance results table below shows the worst-case scenario when one server is hosting all 750 mailboxes. Database I/O Needed disk transfers/sec
120
Database disk transfers/sec
726
Database disk reads/sec
499
Database disk writes/sec
226
Average database disk read latency (ms)
18.6
Average database disk write latency (ms)
1.4
Transaction Log I/O Log disk writes/sec
176
Average log disk write latency (ms)
<1
Database backup / recovery performance This section contains two test reports. The first report measures the sequential read rate of the database files; the second report measures the recovery/replay performance (playing transaction logs into the database). The tests measure the read rates and log replay rates with all mailboxes under concurrent or simultaneous load. Database Read-Only Performance This test measures the maximum backup rate for Volume Shadow Copy Service (VSS) aware backup applications. The following table shows the average and sum of the database reads. Average MB read/sec per database
158
MB read/sec total
632
Transaction Log Recovery / Replay Performance This test measures the maximum rate that log files can play against the databases. The following table shows the average rate for log files played. Each log file is 1 MB in size.
13
Technical white paper | HP MSA 2040 Storage 750 Mailbox Exchange 2013 Mailbox Resiliency Storage Solution
Average number of log files played
503
Average time to play one log file (sec)
2.2
Conclusion The information discussed in this solution brief highlights the tested performance results and configuration best practices for an Exchange solution supporting 750 5GB mailboxes and a 0.16 IOPS workload for each user (200 messages sent and received per user, per day with additional performance and capacity overhead included). The tested solution included an HP MSA 2040 array with eight (8) 2TB 6G SAS 7.2K LFF (3.5-inch) Dual Port MDL drives for the mailbox database storage and transaction logs. Testing demonstrated the storage subsystem provides sufficient IOPS for 750 Exchange users, with each user having a workload up to 0.97 IOPS. The test results and data presented in this document are from the ESRP test framework. Test results are only a guide; customers should test their specific solution to make sure it meets their requirements. The ESRP program is not a benchmarking program. The tests do not measure the maximum throughput for a given solution. Rather, program focus is on producing recommendations from vendors for Exchange Server storage solutions, not direct comparisons among solutions.
14
Technical white paper | HP MSA 2040 Storage 750 Mailbox Exchange 2013 Mailbox Resiliency Storage Solution
Appendix A – stress test Microsoft Exchange Jetstress 2013 Stress Test Result Report Test Summary Overall Test Result Machine Name
Pass WIN-TO4DA3HN869
Test Description Test Start Time
11/19/2013 8:32:25 AM
Test End Time
11/20/2013 8:34:58 AM
Collection Start Time 11/19/2013 8:34:27 AM Collection End Time 11/20/2013 8:34:16 AM Jetstress Version
15.00.0775.000
ESE Version Operating System
15.00.0626.000 Windows Server 2012 Datacenter (6.2.9200.0)
Performance Log
C:\Program Files\Exchange Jetstress\24HR_TEST\Stress_2013_11_19_8_32_34.blg
Database Sizing and Throughput Achieved Transactional I/O per Second 707.084 Target Transactional I/O per Second 120 Initial Database Size (bytes) Final Database Size (bytes)
4034392227840 4055028203520
Database Files (Count)
4
Jetstress System Parameters Thread Count Minimum Database Cache
21 128.0 MB
Maximum Database Cache
1024.0 MB
Insert Operations Delete Operations
40% 20%
Replace Operations
5%
Read Operations
35%
Lazy Commits
70%
Run Background Database Maintenance True Number of Copies per Database
4
Database Configuration Instance3064.1 Log path: E:\ Database: E:\Jetstress001001.edb Instance3064.2 Log path: F:\ Database: F:\Jetstress002001.edb Instance3064.3 Log path: G:\ Database: G:\Jetstress003001.edb Instance3064.4 Log path: H:\ Database: H:\Jetstress004001.edb
Transactional I/O Performance MSExchange I/O I/O Database ==> Database Database Instances Reads Writes Average Average Latency Latency (msec) (msec)
I/O I/O I/O Database Database Database Reads/sec Writes/sec Reads Average Bytes
Instance3064.1 19.155
1.477
121.297
55.488
Instance3064.2 18.486 Instance3064.3 19.195
1.555 1.430
121.247 121.309
55.418 55.464
Instance3064.4 18.385
1.514
121.313
55.549
I/O Database Writes Average Bytes
I/O Log Reads Average Latency (msec)
I/O Log Writes Average Latency (msec)
I/O Log I/O Log I/O Log Reads/sec Writes/sec Reads Average Bytes
I/O Log Writes Average Bytes
33222.975 35462.154 0.000
0.472
0.000
39.258
0.000
7797.800
33232.148 35463.908 0.000 33229.197 35451.881 0.000
0.473 0.472
0.000 0.000
39.253 39.277
0.000 0.000
7791.963 7788.189
33225.750 35454.221 0.000
0.474
0.000
39.278
0.000
7799.116
Background Database Maintenance I/O Performance MSExchange Database ==> Instances Database Maintenance IO Reads/sec Database Maintenance IO Reads Average Bytes Instance3064.1 9.122 261735.230
15
Technical white paper | HP MSA 2040 Storage 750 Mailbox Exchange 2013 Mailbox Resiliency Storage Solution
Instance3064.2
9.126
261736.242
Instance3064.3 Instance3064.4
9.124 9.126
261743.564 261732.403
Log Replication I/O Performance MSExchange Database ==> Instances I/O Log Reads/sec I/O Log Reads Average Bytes Instance3064.1 Instance3064.2
2.664 2.661
229890.912 229930.161
Instance3064.3
2.660
229972.148
Instance3064.4
2.665
229566.428
Total I/O Performance MSExchange I/O Database ==> Database Instances Reads Average Latency (msec) Instance3064.1 19.155
I/O Database Writes Average Latency (msec) 1.477
I/O I/O I/O Database Database Database Reads/sec Writes/sec Reads Average Bytes 130.419
55.488
Instance3064.2 18.486 Instance3064.3 19.195
1.555 1.430
130.373 130.433
55.418 55.464
Instance3064.4 18.385
1.514
130.439
55.549
I/O Database Writes Average Bytes
I/O Log I/O Log I/O Log Writes Reads/sec Writes/sec Average Latency (msec)
I/O Log Reads Average Bytes
49206.377 35462.154 0.588
0.472
2.664
39.258
229890.912 7797.800
49226.815 35463.908 0.586 49214.642 35451.881 0.585
0.473 0.472
2.661 2.660
39.253 39.277
229930.161 7791.963 229972.148 7788.189
49212.814 35454.221 0.585
0.474
2.665
39.278
229566.428 7799.116
Host System Performance Counter
Average
Minimum
Maximum
% Processor Time
1.953
0.470
9.390
Available MBytes
45442.262
45329.000
45519.000
I/O Log Reads Average Latency (msec)
I/O Log Writes Average Bytes
Free System Page Table Entries 16610842.971 16610263.000 16611270.000 Transition Pages RePurposed/sec 0.000 0.000 0.000 Pool Nonpaged Bytes Pool Paged Bytes
48700239.276 43581440.000 54120448.000 87197679.610 79355904.000 100106240.000
Database Page Fault Stalls/sec
0.000
0.000
0.000
Test Log 11/19/2013 8:32:25 AM -- Preparing for testing ... 11/19/2013 8:32:29 AM -- Attaching databases ... 11/19/2013 8:32:29 AM -- Preparations for testing are complete. 11/19/2013 8:32:30 AM -- Starting transaction dispatch .. 11/19/2013 8:32:30 AM -- Database cache settings: (minimum: 128.0 MB, maximum: 1.0 GB) 11/19/2013 8:32:30 AM -- Database flush thresholds: (start: 10.2 MB, stop: 20.5 MB) 11/19/2013 8:32:34 AM -- Database read latency thresholds: (average: 20 msec/read, maximum: 200 msec/read). 11/19/2013 8:32:34 AM -- Log write latency thresholds: (average: 10 msec/write, maximum: 200 msec/write). 11/19/2013 8:32:35 AM -- Operation mix: Sessions 21, Inserts 40%, Deletes 20%, Replaces 5%, Reads 35%, Lazy Commits 70%. 11/19/2013 8:32:35 AM -- Performance logging started (interval: 15000 ms). 11/19/2013 8:32:35 AM -- Attaining prerequisites: 11/19/2013 8:34:27 AM -- \MSExchange Database(JetstressWin)\Database Cache Size, Last: 966684700.0 (lower bound: 966367600.0, upper bound: none) 11/20/2013 8:34:28 AM -- Performance logging has ended. 11/20/2013 8:34:56 AM -- JetInterop batch transaction stats: 366629, 366629, 366628 and 366628. 11/20/2013 8:34:56 AM -- Dispatching transactions ends. 11/20/2013 8:34:56 AM -- Shutting down databases ... 11/20/2013 8:34:58 AM -- Instance3064.1 (complete), Instance3064.2 (complete), Instance3064.3 (complete) and Instance3064.4 (complete) 11/20/2013 8:34:58 AM -- C:\Program Files\Exchange Jetstress\24HR_TEST\Stress_2013_11_19_8_32_34.blg has 5755 samples. 11/20/2013 8:34:58 AM -- Creating test report ... 11/20/2013 8:35:21 AM -- Instance3064.1 has 19.2 for I/O Database Reads Average Latency. 11/20/2013 8:35:21 AM -- Instance3064.1 has 0.5 for I/O Log Writes Average Latency. 11/20/2013 8:35:21 AM -- Instance3064.1 has 0.5 for I/O Log Reads Average Latency. 11/20/2013 8:35:21 AM -- Instance3064.2 has 18.5 for I/O Database Reads Average Latency. 11/20/2013 8:35:21 AM -- Instance3064.2 has 0.5 for I/O Log Writes Average Latency. 11/20/2013 8:35:21 AM -- Instance3064.2 has 0.5 for I/O Log Reads Average Latency. 11/20/2013 8:35:21 AM -- Instance3064.3 has 19.2 for I/O Database Reads Average Latency. 11/20/2013 8:35:21 AM -- Instance3064.3 has 0.5 for I/O Log Writes Average Latency. 11/20/2013 8:35:21 AM -- Instance3064.3 has 0.5 for I/O Log Reads Average Latency. 11/20/2013 8:35:21 AM -- Instance3064.4 has 18.4 for I/O Database Reads Average Latency. 11/20/2013 8:35:21 AM -- Instance3064.4 has 0.5 for I/O Log Writes Average Latency. 11/20/2013 8:35:21 AM -- Instance3064.4 has 0.5 for I/O Log Reads Average Latency. 11/20/2013 8:35:21 AM -- Test has 0 Maximum Database Page Fault Stalls/sec. 11/20/2013 8:35:21 AM -- The test has 0 Database Page Fault Stalls/sec samples higher than 0.
16
Technical white paper | HP MSA 2040 Storage 750 Mailbox Exchange 2013 Mailbox Resiliency Storage Solution
11/20/2013 8:35:21 AM -- C:\Program Files\Exchange Jetstress\24HR_TEST\Stress_2013_11_19_8_32_34.xml has 5747 samples queried.
Appendix B – performance test Microsoft Exchange Jetstress 2013 Performance Test Result Report Test Summary Overall Test Result
Pass
Machine Name Test Description
WIN-TO4DA3HN869
Test Start Time Test End Time
11/18/2013 10:03:52 PM 11/19/2013 6:29:55 AM
Collection Start Time 11/18/2013 10:07:02 PM Collection End Time 11/19/2013 12:07:02 AM Jetstress Version
15.00.0775.000
ESE Version
15.00.0626.000
Operating System
Windows Server 2012 Datacenter (6.2.9200.0)
Performance Log
C:\Program Files\Exchange Jetstress\2HR_TEST\Performance_2013_11_18_22_4_0.blg
Database Sizing and Throughput Achieved Transactional I/O per Second 725.542 Target Transactional I/O per Second Initial Database Size (bytes)
120 4026540490752
Final Database Size (bytes) Database Files (Count)
4034392227840 4
Jetstress System Parameters Thread Count
21
Minimum Database Cache Maximum Database Cache
128.0 MB 1024.0 MB
Insert Operations
40%
Delete Operations Replace Operations
20% 5%
Read Operations
35%
Lazy Commits
70%
Run Background Database Maintenance True Number of Copies per Database
4
Database Configuration Instance928.1 Log path: E:\ Database: E:\Jetstress001001.edb Instance928.2 Log path: F:\ Database: F:\Jetstress002001.edb Instance928.3 Log path: G:\ Database: G:\Jetstress003001.edb Instance928.4 Log path: H:\ Database: H:\Jetstress004001.edb
Transactional I/O Performance MSExchange I/O I/O Database ==> Database Database Instances Reads Writes Average Average Latency Latency (msec) (msec) Instance928.1 19.600 1.427
I/O I/O I/O Database Database Database Reads/sec Writes/sec Reads Average Bytes
I/O Database Writes Average Bytes
I/O Log Reads Average Latency (msec)
I/O Log Writes Average Latency (msec)
I/O Log I/O Log I/O Log Reads/sec Writes/sec Reads Average Bytes
I/O Log Writes Average Bytes
124.924
56.819
33163.704 37392.008 0.000
0.464
0.000
44.044
0.000
7840.436
Instance928.2 17.950 Instance928.3 18.334
1.487 1.353
125.052 124.593
56.720 55.896
33136.425 37454.634 0.000 33137.014 37553.683 0.000
0.466 0.464
0.000 0.000
44.042 43.493
0.000 0.000
7867.102 7840.565
Instance928.4 18.355
1.434
124.834
56.704
33144.532 37437.730 0.000
0.465
0.000
44.006
0.000
7860.530
17
Technical white paper | HP MSA 2040 Storage 750 Mailbox Exchange 2013 Mailbox Resiliency Storage Solution
Background Database Maintenance I/O Performance MSExchange Database ==> Instances Database Maintenance IO Reads/sec Database Maintenance IO Reads Average Bytes Instance928.1
9.112
261686.658
Instance928.2 Instance928.3
9.128 9.129
261692.535 261598.697
Instance928.4
9.126
261642.521
Log Replication I/O Performance MSExchange Database ==> Instances I/O Log Reads/sec I/O Log Reads Average Bytes Instance928.1
3.011
231589.056
Instance928.2
3.030
231588.395
Instance928.3 Instance928.4
2.978 3.023
231589.217 231105.232
Total I/O Performance MSExchange I/O Database ==> Database Instances Reads Average Latency (msec)
I/O I/O I/O Database Database Database Writes Reads/sec Writes/sec Average Latency (msec)
I/O Database Reads Average Bytes
Instance928.1 19.600
1.427
134.036
56.819
Instance928.2 17.950
1.487
134.180
56.720
Instance928.3 18.334
1.353
133.722
Instance928.4 18.355
1.434
133.960
I/O Database Writes Average Bytes
I/O Log I/O Log I/O Log Writes Reads/sec Writes/sec Average Latency (msec)
I/O Log Reads Average Bytes
48699.001 37392.008 0.532
0.464
3.011
44.044
231589.056 7840.436
48684.375 37454.634 0.595
0.466
3.030
44.042
231588.395 7867.102
55.896
48734.297 37553.683 0.582
0.464
2.978
43.493
231589.217 7840.565
56.704
48710.891 37437.730 0.592
0.465
3.023
44.006
231105.232 7860.530
Host System Performance Counter
Average
Minimum
Maximum
% Processor Time Available MBytes
1.985 45439.433
0.652 45376.000
2.720 45464.000
Free System Page Table Entries
16610753.971 16610435.000 16611118.000
I/O Log Reads Average Latency (msec)
I/O Log Writes Average Bytes
Transition Pages RePurposed/sec 0.000 0.000 0.000 Pool Nonpaged Bytes 48504832.000 47394816.000 48828416.000 Pool Paged Bytes
81693201.067 81473536.000 82001920.000
Database Page Fault Stalls/sec
0.000
0.000
0.000
Test Log 11/18/2013 10:03:52 PM -- Preparing for testing ... 11/18/2013 10:03:56 PM -- Attaching databases ... 11/18/2013 10:03:56 PM -- Preparations for testing are complete. 11/18/2013 10:03:56 PM -- Starting transaction dispatch .. 11/18/2013 10:03:56 PM -- Database cache settings: (minimum: 128.0 MB, maximum: 1.0 GB) 11/18/2013 10:03:56 PM -- Database flush thresholds: (start: 10.2 MB, stop: 20.5 MB) 11/18/2013 10:04:00 PM -- Database read latency thresholds: (average: 20 msec/read, maximum: 100 msec/read). 11/18/2013 10:04:00 PM -- Log write latency thresholds: (average: 10 msec/write, maximum: 100 msec/write). 11/18/2013 10:04:02 PM -- Operation mix: Sessions 21, Inserts 40%, Deletes 20%, Replaces 5%, Reads 35%, Lazy Commits 70%. 11/18/2013 10:04:02 PM -- Performance logging started (interval: 15000 ms). 11/18/2013 10:04:02 PM -- Attaining prerequisites: 11/18/2013 10:07:02 PM -- \MSExchange Database(JetstressWin)\Database Cache Size, Last: 978800600.0 (lower bound: 966367600.0, upper bound: none) 11/19/2013 12:07:03 AM -- Performance logging has ended. 11/19/2013 6:29:53 AM -- JetInterop batch transaction stats: 139671, 139671, 139671 and 139670. 11/19/2013 6:29:53 AM -- Dispatching transactions ends. 11/19/2013 6:29:53 AM -- Shutting down databases ... 11/19/2013 6:29:55 AM -- Instance928.1 (complete), Instance928.2 (complete), Instance928.3 (complete) and Instance928.4 (complete) 11/19/2013 6:29:55 AM -- C:\Program Files\Exchange Jetstress\2HR_TEST\Performance_2013_11_18_22_4_0.blg has 491 samples. 11/19/2013 6:29:55 AM -- Creating test report ... 11/19/2013 6:29:58 AM -- Instance928.1 has 19.6 for I/O Database Reads Average Latency. 11/19/2013 6:29:58 AM -- Instance928.1 has 0.5 for I/O Log Writes Average Latency. 11/19/2013 6:29:58 AM -- Instance928.1 has 0.5 for I/O Log Reads Average Latency. 11/19/2013 6:29:58 AM -- Instance928.2 has 17.9 for I/O Database Reads Average Latency. 11/19/2013 6:29:58 AM -- Instance928.2 has 0.5 for I/O Log Writes Average Latency. 11/19/2013 6:29:58 AM -- Instance928.2 has 0.5 for I/O Log Reads Average Latency. 11/19/2013 6:29:58 AM -- Instance928.3 has 18.3 for I/O Database Reads Average Latency. 11/19/2013 6:29:58 AM -- Instance928.3 has 0.5 for I/O Log Writes Average Latency. 11/19/2013 6:29:58 AM -- Instance928.3 has 0.5 for I/O Log Reads Average Latency. 11/19/2013 6:29:58 AM -- Instance928.4 has 18.4 for I/O Database Reads Average Latency. 11/19/2013 6:29:58 AM -- Instance928.4 has 0.5 for I/O Log Writes Average Latency.
18
Technical white paper | HP MSA 2040 Storage 750 Mailbox Exchange 2013 Mailbox Resiliency Storage Solution
11/19/2013 6:29:58 AM -- Instance928.4 has 0.5 for I/O Log Reads Average Latency. 11/19/2013 6:29:58 AM -- Test has 0 Maximum Database Page Fault Stalls/sec. 11/19/2013 6:29:58 AM -- The test has 0 Database Page Fault Stalls/sec samples higher than 0. 11/19/2013 6:29:58 AM -- C:\Program Files\Exchange Jetstress\2HR_TEST\Performance_2013_11_18_22_4_0.xml has 479 samples queried.
Appendix C – streaming backup test Microsoft Exchange Jetstress 2013 Database backup Test Result Report Database Backup Statistics - All Database Instance Database Size (MBytes) Elapsed Backup Time MBytes Transferred/sec Instance2232.1 Instance2232.2
966786.09 966794.09
01:43:15 01:40:40
156.05 160.06
Instance2232.3 Instance2232.4
966778.09 966786.09
01:43:57 01:40:20
155.00 160.59
Avg
157.92
Sum
631.69
Jetstress System Parameters Thread Count 21 Minimum Database Cache 128.0 MB Maximum Database Cache 1024.0 MB Insert Operations 40% Delete Operations
20%
Replace Operations Read Operations
5% 35%
Lazy Commits
70%
Database Configuration Instance2232.1 Log path: E:\ Database: E:\Jetstress001001.edb Instance2232.2 Log path: F:\ Database: F:\Jetstress002001.edb Instance2232.3 Log path: G:\ Database: G:\Jetstress003001.edb Instance2232.4 Log path: H:\ Database: H:\Jetstress004001.edb
Transactional I/O Performance MSExchange I/O I/O Database ==> Database Database Instances Reads Writes Average Average Latency Latency (msec) (msec) Instance2232.1 1.848 0.000
I/O I/O I/O Database Database Database Reads Reads/sec Writes/sec Average Bytes
I/O Database Writes Average Bytes
I/O Log Reads Average Latency (msec)
I/O Log Writes Average Latency (msec)
I/O Log I/O Log I/O Log Reads/sec Writes/sec Reads Average Bytes
I/O Log Writes Average Bytes
624.050
0.000
262144.000 0.000
0.000
0.000
0.000
0.000
0.000
0.000
Instance2232.2 2.840
0.000
640.034
0.000
262144.000 0.000
0.000
0.000
0.000
0.000
0.000
0.000
Instance2232.3 2.918
0.000
619.879
0.000
262144.000 0.000
0.000
0.000
0.000
0.000
0.000
0.000
Instance2232.4 2.811
0.000
642.249
0.000
262144.000 0.000
0.000
0.000
0.000
0.000
0.000
0.000
Host System Performance Counter
Average
Minimum
Maximum
% Processor Time Available MBytes
6.335 46574.092
2.297 46558.000
8.532 46578.000
Free System Page Table Entries 16611316.063 16610559.000 16611734.000 Transition Pages RePurposed/sec 0.000 0.000 0.000 Pool Nonpaged Bytes
41616171.285 40038400.000 41885696.000
Pool Paged Bytes Database Page Fault Stalls/sec
81787582.454 78589952.000 86679552.000 0.000 0.000 0.000
19
Technical white paper | HP MSA 2040 Storage 750 Mailbox Exchange 2013 Mailbox Resiliency Storage Solution
Test Log 11/20/2013 10:11:46 AM -- Preparing for testing ... 11/20/2013 10:11:50 AM -- Attaching databases ... 11/20/2013 10:11:50 AM -- Preparations for testing are complete. 11/20/2013 10:11:55 AM -- Performance logging started (interval: 30000 ms). 11/20/2013 10:11:55 AM -- Backing up databases ... 11/20/2013 11:55:53 AM -- Performance logging has ended. 11/20/2013 11:55:53 AM -- Instance2232.1 (100% processed), Instance2232.2 (100% processed), Instance2232.3 (100% processed) and Instance2232.4 (100% processed) 11/20/2013 11:55:53 AM -- C:\Program Files\Exchange Jetstress\DBBACKUP_TEST\DatabaseBackup_2013_11_20_10_11_50.blg has 207 samples. 11/20/2013 11:55:53 AM -- Creating test report ...
Appendix D – soft recovery test Microsoft Exchange Jetstress 2013 SoftRecovery Test Result Report Soft-Recovery Statistics - All Database Instance Log files replayed Elapsed seconds Instance2236.1 504 1103.7300958 Instance2236.2
503
1088.2074849
Instance2236.3 Instance2236.4
501 506
1113.0903151 1086.1235552
Avg Sum
503 2014
1097.788 4391.151451
Database Configuration Instance2236.1 Log path: E:\ Database: E:\Jetstress001001.edb Instance2236.2 Log path: F:\ Database: F:\Jetstress002001.edb Instance2236.3 Log path: G:\ Database: G:\Jetstress003001.edb Instance2236.4 Log path: H:\ Database: H:\Jetstress004001.edb
Transactional I/O Performance MSExchange I/O I/O I/O I/O Database ==> Database Database Database Database Instances Reads Writes Reads/sec Writes/sec Average Average Latency Latency (msec) (msec)
I/O Database Reads Average Bytes
Instance2236.1 22.079
0.413
373.908
1.798
Instance2236.2 21.561 Instance2236.3 22.230
0.418 0.374
380.904 371.701
1.844 1.782
Instance2236.4 21.595
0.430
381.972
1.858
I/O Database Writes Average Bytes
I/O Log Reads Average Latency (msec)
I/O Log I/O Log I/O Log Writes Reads/sec Writes/sec Average Latency (msec)
I/O Log Reads Average Bytes
I/O Log Writes Average Bytes
40172.872 28732.235 3.030
0.000
2.250
0.000
182477.176 0.000
40232.982 28923.709 2.734 40283.919 27873.690 2.589
0.000 0.000
2.307 2.229
0.000 0.000
184764.357 0.000 178009.623 0.000
40252.552 28977.672 2.747
0.000
2.334
0.002
185128.364 7.642
Background Database Maintenance I/O Performance MSExchange Database ==> Instances Database Maintenance IO Reads/sec Database Maintenance IO Reads Average Bytes Instance2236.1 0.000 0.000 Instance2236.2 Instance2236.3
0.000 0.000
0.000 0.000
Instance2236.4
0.000
0.000
Total I/O Performance MSExchange I/O Database ==> Database Instances Reads Average Latency (msec) Instance2236.1 22.079
20
I/O Database Writes Average Latency (msec) 0.413
I/O I/O I/O Database Database Database Reads/sec Writes/sec Reads Average Bytes 373.908
1.798
I/O Database Writes Average Bytes
I/O Log Reads Average Latency (msec)
40172.872 28732.235 3.030
I/O Log I/O Log I/O Log Writes Reads/sec Writes/sec Average Latency (msec)
I/O Log Reads Average Bytes
0.000
182477.176 0.000
2.250
0.000
I/O Log Writes Average Bytes
Technical white paper | HP MSA 2040 Storage 750 Mailbox Exchange 2013 Mailbox Resiliency Storage Solution
Instance2236.2 21.561
0.418
380.904
1.844
40232.982 28923.709 2.734
0.000
2.307
0.000
184764.357 0.000
Instance2236.3 22.230 Instance2236.4 21.595
0.374 0.430
371.701 381.972
1.782 1.858
40283.919 27873.690 2.589 40252.552 28977.672 2.747
0.000 0.000
2.229 2.334
0.000 0.002
178009.623 0.000 185128.364 7.642
Host System Performance Counter
Average
Minimum
Maximum
% Processor Time Available MBytes
3.366 45510.149
0.904 45434.000
15.352 46527.000
Free System Page Table Entries
16611359.653 16611013.000 16611745.000
Transition Pages RePurposed/sec 0.000
0.000
0.000
Pool Nonpaged Bytes
51172288.698 51126272.000 51240960.000
Pool Paged Bytes
80005894.516 79540224.000 81379328.000
Database Page Fault Stalls/sec
0.000
0.000
0.000
Test Log 11/20/2013 11:59:37 AM -- Preparing for testing ... 11/20/2013 11:59:41 AM -- Attaching databases ... 11/20/2013 11:59:41 AM -- Preparations for testing are complete. 11/20/2013 11:59:41 AM -- Starting transaction dispatch ... 11/20/2013 11:59:41 AM -- Database cache settings: (minimum: 128.0 MB, maximum: 1.0 GB) 11/20/2013 11:59:41 AM -- Database flush thresholds: (start: 10.2 MB, stop: 20.5 MB) 11/20/2013 11:59:46 AM -- Database read latency thresholds: (average: 20 msec/read, maximum: 100 msec/read). 11/20/2013 11:59:46 AM -- Log write latency thresholds: (average: 10 msec/write, maximum: 100 msec/write). 11/20/2013 11:59:46 AM -- Operation mix: Sessions 21, Inserts 40%, Deletes 20%, Replaces 5%, Reads 35%, Lazy Commits 70%. 11/20/2013 11:59:46 AM -- Performance logging started (interval: 15000 ms). 11/20/2013 11:59:46 AM -- Generating log files ... 11/20/2013 1:22:22 PM -- E:\ (100.8% generated), F:\ (100.6% generated), G:\ (100.2% generated) and H:\ (101.2% generated) 11/20/2013 1:22:23 PM -- Performance logging has ended. 11/20/2013 1:22:23 PM -- JetInterop batch transaction stats: 21387, 21387, 21387 and 21387. 11/20/2013 1:22:23 PM -- Dispatching transactions ends. 11/20/2013 1:22:23 PM -- Shutting down databases ... 11/20/2013 1:22:24 PM -- Instance2236.1 (complete), Instance2236.2 (complete), Instance2236.3 (complete) and Instance2236.4 (complete) 11/20/2013 1:22:24 PM -- C:\Program Files\Exchange Jetstress\DBRESTORE_TEST\Performance_2013_11_20_11_59_46.blg has 329 samples. 11/20/2013 1:22:24 PM -- Creating test report ... 11/20/2013 1:22:26 PM -- Instance2236.1 has 20.1 for I/O Database Reads Average Latency. 11/20/2013 1:22:26 PM -- Instance2236.1 has 0.5 for I/O Log Writes Average Latency. 11/20/2013 1:22:26 PM -- Instance2236.1 has 0.5 for I/O Log Reads Average Latency. 11/20/2013 1:22:26 PM -- Instance2236.2 has 19.6 for I/O Database Reads Average Latency. 11/20/2013 1:22:26 PM -- Instance2236.2 has 0.5 for I/O Log Writes Average Latency. 11/20/2013 1:22:26 PM -- Instance2236.2 has 0.5 for I/O Log Reads Average Latency. 11/20/2013 1:22:26 PM -- Instance2236.3 has 20.4 for I/O Database Reads Average Latency. 11/20/2013 1:22:26 PM -- Instance2236.3 has 0.5 for I/O Log Writes Average Latency. 11/20/2013 1:22:26 PM -- Instance2236.3 has 0.5 for I/O Log Reads Average Latency. 11/20/2013 1:22:26 PM -- Instance2236.4 has 19.6 for I/O Database Reads Average Latency. 11/20/2013 1:22:26 PM -- Instance2236.4 has 0.5 for I/O Log Writes Average Latency. 11/20/2013 1:22:26 PM -- Instance2236.4 has 0.5 for I/O Log Reads Average Latency. 11/20/2013 1:22:26 PM -- Test has 0 Maximum Database Page Fault Stalls/sec. 11/20/2013 1:22:26 PM -- The test has 0 Database Page Fault Stalls/sec samples higher than 0. 11/20/2013 1:22:26 PM -- C:\Program Files\Exchange Jetstress\DBRESTORE_TEST\Performance_2013_11_20_11_59_46.xml has 328 samples queried. 11/20/2013 1:22:26 PM -- C:\Program Files\Exchange Jetstress\DBRESTORE_TEST\Performance_2013_11_20_11_59_46.html was saved. 11/20/2013 2:12:19 PM -- Performance logging started (interval: 2000 ms). 11/20/2013 2:12:19 PM -- Recovering databases ... 11/20/2013 2:30:53 PM -- Performance logging has ended. 11/20/2013 2:30:53 PM -- Instance2236.1 (1103.7300958), Instance2236.2 (1088.2074849), Instance2236.3 (1113.0903151) and Instance2236.4 (1086.1235552) 11/20/2013 2:30:53 PM -- C:\Program Files\Exchange Jetstress\DBRESTORE_TEST\SoftRecovery_2013_11_20_14_12_17.blg has 550 samples. 11/20/2013 2:30:53 PM -- Creating test report ...
21
Technical white paper | HP MSA 2040 Storage 750 Mailbox Exchange 2013 Mailbox Resiliency Storage Solution with Microsoft Hyper-V
For more information HP MSA 2040 SAN Storage, hp.com/go/msa2040 HP ProLiant DL380p Gen8 Server, hp.com/servers/dl380 HP StoreFabric Storage Networking, hp.com/go/storefabric To read more about HP solutions for Microsoft Exchange, see: hp.com/solutions/exchange For general information on Exchange sizing and best practices when deploying with HP solutions, see: hp.com/solutions/activeanswers/exchange For more information on Microsoft’s Exchange Solution Reviewed Program (ESRP), see: technet.microsoft.com/en-us/exchange/ff182054.aspx
To help us improve our documents, provide feedback at hp.com/solutions/feedback.
Sign up for updates hp.com/go/getupdated © Copyright 2014 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice. The only warranties for HP products and services are set forth in the express warranty statements accompanying such products and services. Nothing herein should be construed as constituting an additional warranty. HP shall not be liable for technical or editorial errors or omissions contained herein. Microsoft and Windows are U.S. registered trademarks of Microsoft Corporation. Intel and Xeon are trademarks of Intel Corporation in the U.S. and other countries. ENERGY STAR is a registered mark owned by the U.S. government. 4AA5-0800ENW, January 2014
