Technical white paper HP D2600 disk enclosure and HP Smart Array P411 2,000 user 4 GB mailbox resiliency Exchange 2010 storage solution Table of contents Overview 2 Disclaimer 2 Features of the tested solution 2 HP D2000 disk enclosure family 2 Solution description 3 Storage configuration 4 Targeted customer profile 6 Tested deployment 6 Best practices 8 Test results summary 8 Conclusion 10 Appendix A Test reports 11 2 hour test 11 24 hour test 13 backup test 15 Soft recovery test 16 For more information 18
Overview This solution brief provides information on an HP D2600 disk enclosure and HP Smart Array P411 storage solution for Microsoft Exchange Server 2010 (Exchange), based on the Microsoft Exchange Solution Review Program (ESRP) Storage Program. The ESRP Storage Program was developed by Microsoft Corporation to provide a common storage testing framework for vendors to provide information on their storage solutions for Microsoft Exchange Server software. For more details on the Microsoft ESRP V3.0 Storage Program, please see: http://technet.microsoft.com/en-us/exchange/2010/ff182054.aspx Tested with: ESRP Storage Version 3.0 Test Date: June 2012 Disclaimer This document has been produced 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 implements the Availability Group (DAG) mailbox resiliency features in Exchange Server 2010. The DAG architecture allows two or more mailbox servers (up to 16) to be grouped together for increased database availability. Mailbox servers participating in the DAG manage mailbox database replications and configuration of database copies. DAG members also monitor database operations and in the event of a database, server, storage or network failure, can initiate an automatic database-level recovery, by activating a (passive) database copy on another server. This solution brief describes a tested configuration for deploying Microsoft Exchange Server 2010 (Exchange) in a 2,000 user mailbox resiliency environment. The performance results and best practices outlined in this document provide tested guidelines for configuring Exchange using the HP D2600 disk enclosure, HP Smart Array P411 controller and ProLiant DL360 G7 servers in a direct attached storage (DAS) configuration. For this solution, a single DAG was designed for supporting 2,000 mailboxes. The mailboxes were 4 GB in size and evenly distributed across 6 active databases. To provide mailbox resiliency, there are two copies of each database (active and passive). The active and passive database copies are hosted on separate server and storage configurations. With no server failures, each server hosts 3 active databases. With one of the servers offline (due to hardware failure, maintenance, power outage, etc.), the other server would host 6 active databases. This solution was tested with all active mailbox databases on one server, simulating the worst case scenario. The HP D2000 disk enclosure family offers affordable, modular and flexible storage, as detailed below. HP D2000 disk enclosure family Manage your small and midrange business growing storage needs by deploying the HP D2000 disk enclosures, the next generation 6Gb SAS low-cost, flexible, and tiered external storage system. The HP D2000 disk enclosures are ideal for small application environments in small and medium businesses, remote offices and departmental locations. The 6Gb SAS enclosures Large Form Factor (LFF) D2600 with 12 drive bays and Small Form Factor (SFF) D2700 with 25 drive bays offer modular solutions to simplify capacity expansion of HP ProLiant and Integrity server environments to external storage without having to make the full move to SAN or NAS. HP MSA storage (SFF only) can also be expanded with the D2700. This allows you to buy what is needed today and purchase additional capacity as data storage needs grow. 2
Affordable storage HP D2000 disk enclosures double the transfer rate of current 3Gb solutions, adding crucial bandwidth. The HP D2000 disk enclosures modular platform allows customers to purchase based on current need and future growth, no hidden management charges. Ability to add enterprise-class dual-port SAS drives as the need and budget dictates Affordable redundancy and reliability, reducing or eliminating single points of failure within the storage network. Flexible and scalable Flexibility to mix and match SAS and SATA drives in the same enclosure. Extends benefit of tiered storage to direct attach customers. Solid State drives supported in the D2700. The D2600 LFF can grow up to 96 drives. Up to four Large Form Factor (LFF) enclosures can be connected together from a single controller port. The D2700 SFF can grow up to 200 drives. Up to four Small Form Factor (SFF) enclosures can be connected together from a single controller port. Easy to manage The HP D2000 disk enclosures use familiar HP ProLiant management tools including HP Systems Insight Manager and the HP Array Configuration Utility (ACU) software. The HP D2000 disk enclosures are compatible with selected HP Smart Array SAS controllers. Little IT expertise is required. Solution description This solution brief provides tested best practices to help administrators implement the Exchange deployment, with the following criteria: 2,000 users total, with (2) mailbox servers 0.24 IOPS per user (0.2 with additional 20% headroom included) 4 GB tested mailbox size Availability Group (DAG) deployment with 2 copies of each database in the DAG 6 databases total on each server (3 active and 3 passive) during normal operations background maintenance (DBM) configured during testing Storage in this solution is deployed as RAID 1 pairs. See the Storage configuration section for more information. HP has developed the HP Sizer for Microsoft Exchange Server 2010 to assist customers with proper server and storage sizing for their Exchange Server 2010 deployments. The algorithms developed and implemented in this tool are based upon extensive testing of Exchange Server 2010 server roles and technologies, HP s deep experience with Microsoft Exchange, and intricate knowledge of server CPU, memory and subsystems. This tool is free to download and use, and is designed to simplify configuration across a broad range of deployments. For more information, please visit: hp.com/solutions/microsoft/exchange2010/sizer 3
1 4 7 10 3 6 9 12 1 2 Figure 1 below illustrates the tested solution hardware. Each server in this solution utilizes all (12) 3TB MDL SAS Disks for database, logs and content indexing all on the same volumes. Recovery space (often referred as the restore LUN) can be provided using drives within the server. Figure 1: Rack mounted view of the hardware used in this solution 1 6G DP 15k SAS 146 GB 5 1 6G DP 15k SAS 146 GB 2 6G DP 15k SAS 146 GB 3 6G DP 15k SAS 146 GB 8 4 UID SID HP ProLiant DL360 G7 HP StorageWorks D2600 2 UID 1. Host (DL360 G7) 2. Storage (D2600) Storage configuration In this solution, the D2600 disk enclosure is attached to a Smart Array P411 controller and configured using the HP Array Configuration Utility (ACU). The Smart Array P411 controller manages RAID configuration and LUN ownership. For storage performance and redundancy, the D2600 and P411 are cabled as a dual-domain configuration. The dual-domain configuration uses two independent SAS data paths from the servers to the storage. The dual-domain communication path prevents disk from being interrupted in the event of a single data path failure. For more detail on implementing a dual-domain configuration with the D2600, please see: http://h20000.www2.hp.com/bizsupport/techsupport/home.jsp?prodtypeid=12169&prodseriesid=3954787 Figure 2 below illustrates the dual-domain configuration used in this solution. Figure 2: Dual-Domain Configuration 1 Smart Array P411 ilo UID PLC B 1200W 94% TOP PLC B 1200W 94% UID 2 3 PS 1 A Mfg P1 P2 UID Attention This module fits 3.5" large form factor drives only. PS 2 1 B 4 Mfg P1 P2 UID Attention This module fits 3.5" large form factor drives only. 2 1. Host (DL360 G7) 2. Storage (D2600) 3. Primary Path 4. Alternate Path The disk layout containing the mailbox databases and logs were created using RAID 1. Each RAID 1 disk group consisted of (2) 3 TB disks, for a total of (6) RAID groups. Since Exchange was configured for multiple database copies, the databases and logs resided on the same volumes. 4
Figures 3 and 4 below show how the active mailbox databases would be distributed during normal and failed operations in order to maintain production. This solution was tested will all active mailbox databases on one server, simulating the worst case scenario. Figure 3: Primary Data Center with (2) Member DAG Normal Operation Active MBX1 DAG1 MBX2 Passive DB1 DB2 DB3 DB4 DB5 DB6 DB1 DB2 DB3 DB4 DB5 DB6 Figure 4: Primary Data Center with (2) Member DAG MBX2 Offline Due to Fault Active MBX1 DAG1 MBX2 Passive DB1 DB2 DB3 DB4 DB5 DB6 DB1 DB2 DB3 DB4 DB5 DB6 5
Targeted customer profile This solution brief provides tested best practices to help administrators implement the Exchange deployment, with the following criteria: 2000 users total, with (2) active Exchange servers and two (2) database copies 0.24 IOPS per user (0.2 with additional 20% headroom) 4 GB tested mailbox size Availability Group (DAG) deployment with 2 copies of each database in the DAG 6 databases total on each server (3 active and 3 passive) during normal operations background maintenance (DBM) configured during testing Tested deployment The following table summarizes the testing environment. Simulated Exchange configuration Number of Exchange mailboxes simulated 2000 Number of Availability Groups (DAGs) 1 Number of mailbox servers/dag 2 Number of active mailboxes/server 1000 Number of databases/host 6 Number of copies/database 2 Number of mailboxes/database 167 Simulated profile: operations per second per mailbox (IOPS), includes 20% headroom maintenance (DBM) + Log LUN size Total database size for performance testing 0.24 Tested with 24/7 DBM enabled 3 TB 8.64 TB (6 databases * 1.44 TB) % storage capacity used by Exchange database 48% (8.64 TB / 18 TB) 6
Storage hardware Storage Connectivity Storage model and OS/firmware version Storage cache Number of storage controllers Number of storage ports Maximum bandwidth of storage connectivity to host Storage Controller Controller Cache Host server type SAS HP D2600 (0134 firmware) See below 1 per server Connected (2) 4x 6Gb SAS connectors 24 Gb per 4x SAS connector (2 in use for this solution) HP Smart Array P411 1 GB Flash-Backed Write Cache (25/75 read/write = default) HP ProLiant DL360 G7 Total number of disks tested in solution 12 Maximum number of spindles that can be hosted in the storage 96 (12 per D2600 x8) Storage software Smart Array P411 Driver 6.24.0.64 Queue depth & target setting Multipathing Host OS Not applicable Not applicable Windows Server 2008 R2 Enterprise with SP1 ESE.dll file version 14.02.0283.000 Storage disk configuration (mailbox & log disks) Disk type, speed, and firmware revision Raw capacity per disk (TB) HP 3 TB 6Gb SAS 7.2k RPM LFF 3.5 MDL HDD (firmware: HPD3) 3 TB Number of physical disks in test 12 Total raw storage capacity (TB) Disk slice size (GB) RAID level Total formatted capacity (TB) 36 TB Not applicable RAID 1 (2 disks per RAID group) 18 TB Storage capacity utilization 50% capacity utilization 24% Formatted capacity / Total raw capacity Total size / Total raw capacity * Capacity of drives, and therefore the raw data capacity of the disk array, is based on 1K = 1000, not 1024. This means that 1 GB = 1,000,000,000 bytes, not (1,073,741,824 bytes) 7
Best practices Exchange Server is a disk-intensive application. For Exchange 2010 best practices on storage design from Microsoft, visit http://technet.microsoft.com/en-us/library/dd346703.aspx. For Exchange Server solution guidance, visit HP at hp.com/solutions/microsoft/exchange. The best practices for this solution are: Drivers and Firmware Ensure firmware and drivers for all hardware are up to date. For more information, see: http://support.hp.com Physical Drive Write Cache Unlike controller-level cache, individual drive write caches are not protected. If the server loses power or if a hot plug drive is physically removed from the array during operations, unwritten data in the physical drives write caches will be lost. HP recommends Physical Drive Write Cache be disabled on the HP Smart Array P411 controller. For more information, see: http://h20000.www2.hp.com/bc/docs/support/supportmanual/c02278484/c02278484.pdf Accelerator Ratio and Flash-Backed Write Cache (FBWC) HP recommends setting the Accelerator Ratio to 75 percent write cache and 25 percent read cache (battery-backed cache). 1 GB or more of Flash-Backed Write Cache (FBWC) is recommended. Dual Domain The HP Smart Array Advanced Pack is required to enable this feature. Dual domain creates redundant pathways for external drives from server to storage devices. The redundant paths created by these configurations reduce or eliminate single points of failure within the storage network. For more information on the Smart Array Advanced Pack, see: http://h18004.www1.hp.com/products/quickspecs/13200_div/13200_div.pdf Jetstress 2010 Before installing Exchange Server 2010, test your storage solution with Jetstress to ensure it is configured properly and meets your requirements. For more information on Jetstress, see: http://technet.microsoft.com/en-us/library/ff706601 Mailbox Copies This solution was tested with 2 copies of the Exchange mailbox database. When using a JBOD configuration without RAID storage protection, ensure that there are at least three copies for each Exchange mailbox database in the solution. RAID Level The disks used in this solution were configured as RAID 1 pairs with a RAID stripe size of 256 KB (256 KB or better is recommended). Provisioning of other RAID types are supported but not recommended for this Exchange Server solution. Test results summary This section provides a high-level summary of the test data from ESRP and links to the detailed HTML reports generated by the ESRP testing framework. Click on the underlined headings below to view the HTML report for each test. An HP ProLiant DL360 G7 server was used to drive the Jetstress load, and simulate the storage. For solutions that utilize a uniform building block style where each building block is independent and isolated, ESRP requires that only one of the building blocks be tested. In this case, a single unit was tested as the building block. However, for production deployments of this solution, 2 identical units are required. For site resiliency, a third copy could be added later. 8
Reliability One of the tests in the framework is designed to test reliability over a 24-hour test period. The goal is to verify that the storage can handle high load for a long period of time. Both log and database files will be analyzed for integrity after the stress test to ensure no database/log corruption. Below is an initial list of questions used to validate testing. Q. Were any errors reported in the saved event log file? A. No. There were no relevant errors reported in the event log for the storage reliability testing. Q. Were any errors reported during the database and log checksum process? A. No. There were no errors reported for the database and log checksum process. Storage performance results The Primary Storage performance testing is designed to exercise the storage with maximum sustainable Exchange type of for 2 hours. The test shows how long it takes the storage to respond to an under load. The data below is the sum of all logical disk s and the average of all the logical disk latency in the 2 hour test duration. This solution is designed around a 2 server DAG with a single server failure service level. It is important to understand not only the performance of the storage solution in normal operations with all servers hosting Exchange mailbox databases, but also with a single server failure where the remaining server is running all Exchange mailbox databases. The following table provides performance information for this scenario. Disk Transfers/sec 701 Disk /sec 438 Disk /sec 263 Disk Read (ms) 15.02 Disk Write (ms) 0.84 Transaction Log Log Disk /sec 240 Log Disk Write (ms) 0.17 backup/recovery performance There are two test reports in this section. The first one is to measure the sequential read rate of the database files, and the second is to measure the recovery/replay performance (playing transaction logs into the database). One important characteristic of the default test is that these read rates and log replay rates are measured with all databases under concurrent or simultaneous load. This test case was also run with a single server failure where the remaining server is running all Exchange mailbox databases. read-only performance This test is to measure the maximum rate at which databases could be backed up via a Microsoft Volume Shadow Copy Service (VSS) aware backup application. The following table shows the average rate for a single database file and for the entire server. MB read/sec per database 39.49 MB read/sec per server 226.56 9
Transaction log recovery / replay performance This test is to measure the maximum rate at which the log files can be played against the databases. The following table shows the average rate for 500 log files played in a single storage group. Each log file is 1 MB in size. time to play one log file (sec) 3.61 Conclusion The information discussed in this solution brief highlights the tested performance results and configuration best practices for an Exchange Server solution supporting 2,000 Exchange users with 4 GB mailboxes. Testing was performed with an HP Smart Array P411 controller and D2600 disk enclosure with (12) 3 TB MDL SAS 7.2k RPM disks for the Exchange database storage and transaction logs. The testing demonstrated that the storage subsystem was capable of supporting both the IOPS needed to support 2,000 Exchange users as well as providing additional headroom. This document was developed by HP Storage, and reviewed by the Microsoft Exchange product team. The test results and data presented in this document are based on the tests introduced in the ESRP test framework and should be used as a guide. Each specific deployment should be tested to ensure that their solution performs to the requirements of their specific environment. The ESRP program is not designed to be a benchmarking program. The tests are not designed to measure the maximum throughput for a given solution. Rather, it is focused on producing recommendations from vendors for Exchange Server storage solutions. So the data presented in this document should not be used for direct comparisons among solutions. 10
Appendix A Test reports 2 hour test Test Summary Overall Test Result Machine Name Test Description Test Start Time Test End Time Pass D2600 2000 users, 0.24 IOPS, 4GB Mailbox, Thread Count=4 6/20/2012 2:14:39 PM 6/21/2012 7:58:02 AM Collection Start Time 6/20/2012 2:17:52 PM Collection End Time 6/20/2012 4:17:45 PM Jetstress Version 14.01.0225.017 ESE Version 14.02.0283.000 Operating System Windows Server 2008 R2 Enterprise Service Pack 1 (6.1.7601.65536) Performance Log C:\Jetstress tests\2000 user 4 thread\performance_2012_6_20_14_14_52.blg Sizing and Throughput Achieved Transactional per Second 701.47 Target Transactional per Second 480 Initial Size (bytes) 8597483683840 Final Size (bytes) 8616081227776 Files (Count) 6 Jetstress System Parameters Thread Count Minimum Cache Maximum Cache Insert Operations 40% Delete Operations 20% Replace Operations 5% Read Operations 35% Lazy Commits 70% Run Background Maintenance True Number of Copies per 2 4 (per database) 192.0 MB 1536.0 MB Configuration Instance4956.1 Log path: E:\dblg1 : E:\dblg1\Jetstress001001.edb Instance4956.2 Log path: F:\dblg2 : F:\dblg2\Jetstress002001.edb Instance4956.3 Log path: G:\dblg3 : G:\dblg3\Jetstress003001.edb Instance4956.4 Log path: H:\dblg4 : H:\dblg4\Jetstress004001.edb Instance4956.5 Log path: I:\dblg5 : I:\dblg5\Jetstress005001.edb Instance4956.6 Log path: J:\dblg6 : J:\dblg6\Jetstress006001.edb Transactional Performance MSExchange ==> Instances /sec /sec Log Log Log Log /sec /sec Log Log Instance4956.1 19.184 1.126 73.681 44.141 33950.240 36202.042 0.000 0.179 0.000 40.362 0.000 4449.584 Instance4956.2 14.362 1.019 72.532 43.507 34492.805 36201.516 0.000 0.173 0.000 39.711 0.000 4525.801 Instance4956.3 14.120 0.912 71.925 43.072 34530.017 36214.553 0.000 0.171 0.000 39.550 0.000 4524.763 Instance4956.4 14.340 0.801 73.573 44.205 34503.301 36115.821 0.000 0.168 0.000 39.874 0.000 4490.570 Instance4956.5 13.953 0.684 73.184 43.943 34478.125 36113.834 0.000 0.172 0.000 40.111 0.000 4495.380 Instance4956.6 14.180 0.526 73.533 44.174 34496.028 36153.090 0.000 0.167 0.000 40.297 0.000 4470.200 Background Maintenance Performance MSExchange ==> Instances Maintenance IO /sec Maintenance IO Instance4956.1 24.682 261821.157 Instance4956.2 26.274 261867.887 Instance4956.3 27.991 261877.541 Instance4956.4 26.136 261830.558 Instance4956.5 27.947 261913.182 Instance4956.6 26.287 261880.308 11
Log Replication Performance MSExchange ==> Instances Log /sec Log Instance4956.1 0.725 229163.171 Instance4956.2 0.725 227221.111 Instance4956.3 0.721 226735.595 Instance4956.4 0.721 221394.928 Instance4956.5 0.727 224793.535 Instance4956.6 0.726 227221.111 Total Performance MSExchange ==> Instances /sec /sec Log Log Log Log Log /sec /sec Log Instance4956.1 19.184 1.126 98.362 44.141 91128.760 36202.042 0.337 0.179 0.725 40.362 229163.171 4449.584 Instance4956.2 14.362 1.019 98.806 43.507 94954.766 36201.516 0.673 0.173 0.725 39.711 227221.111 4525.801 Instance4956.3 14.120 0.912 99.917 43.072 98220.958 36214.553 0.332 0.171 0.721 39.550 226735.595 4524.763 Instance4956.4 14.340 0.801 99.709 44.205 94090.906 36115.821 0.432 0.168 0.721 39.874 221394.928 4490.570 Instance4956.5 13.953 0.684 101.131 43.943 97328.180 36113.834 0.342 0.172 0.727 40.111 224793.535 4495.380 Instance4956.6 14.180 0.526 99.820 44.174 94376.200 36153.090 0.702 0.167 0.726 40.297 227221.111 4470.200 Host System Performance Counter Minimum Maximum % Processor Time 0.606 0.381 1.593 Available M 44663.404 44565.000 44798.000 Free System Page Table Entries 33555094.004 33555068.000 33555604.000 Transition Pages RePurposed/sec 0.000 0.000 0.000 Pool Nonpaged 184358510.933 184238080.000 184713216.000 Pool Paged 103770777.600 99627008.000 147685376.000 Page Fault Stalls/sec 0.000 0.000 0.000 Test Log6/20/2012 2:14:39 PM -- Jetstress testing begins... 6/20/2012 2:14:39 PM -- Preparing for testing... 6/20/2012 2:14:46 PM -- Attaching databases... 6/20/2012 2:14:46 PM -- Preparations for testing are complete. 6/20/2012 2:14:46 PM -- Starting transaction dispatch.. 6/20/2012 2:14:46 PM -- cache settings: (minimum: 192.0 MB, maximum: 1.5 GB) 6/20/2012 2:14:46 PM -- flush thresholds: (start: 15.3 MB, stop: 30.7 MB) 6/20/2012 2:14:52 PM -- read latency thresholds: (average: 20 msec/read, maximum: 100 msec/read). 6/20/2012 2:14:52 PM -- Log write latency thresholds: (average: 10 msec/write, maximum: 100 msec/write). 6/20/2012 2:14:56 PM -- Operation mix: Sessions 4, Inserts 40%, Deletes 20%, Replaces 5%, 35%, Lazy Commits 70%. 6/20/2012 2:14:56 PM -- Performance logging started (interval: 15000 ms). 6/20/2012 2:14:56 PM -- Attaining prerequisites: 6/20/2012 2:17:52 PM -- \MSExchange (JetstressWin)\ Cache Size, Last: 1450156000.0 (lower bound: 1449551000.0, upper bound: none) 6/20/2012 4:17:52 PM -- Performance logging has ended. 6/21/2012 7:58:00 AM -- JetInterop batch transaction stats: 219151, 220288, 219390, 219657, 219870 and 219728. 6/21/2012 7:58:00 AM -- Dispatching transactions ends. 6/21/2012 7:58:00 AM -- Shutting down databases... 6/21/2012 7:58:02 AM -- Instance4956.1 (complete), Instance4956.2 (complete), Instance4956.3 (complete), Instance4956.4 (complete), Instance4956.5 (complete) and Instance4956.6 (complete) 6/21/2012 7:58:02 AM -- C:\Jetstress tests\2000 user 4 threads\performance_2012_6_20_14_14_52.blg has 491 samples. 6/21/2012 7:58:02 AM -- Creating test report... 6/21/2012 7:58:05 AM -- Instance4956.1 has 19.2 for. 6/21/2012 7:58:05 AM -- Instance4956.1 has 0.2 for Log. 6/21/2012 7:58:05 AM -- Instance4956.1 has 0.2 for Log. 6/21/2012 7:58:05 AM -- Instance4956.2 has 14.4 for. 6/21/2012 7:58:05 AM -- Instance4956.2 has 0.2 for Log. 6/21/2012 7:58:05 AM -- Instance4956.2 has 0.2 for Log. 6/21/2012 7:58:05 AM -- Instance4956.3 has 14.1 for. 6/21/2012 7:58:05 AM -- Instance4956.3 has 0.2 for Log. 6/21/2012 7:58:05 AM -- Instance4956.3 has 0.2 for Log. 6/21/2012 7:58:05 AM -- Instance4956.4 has 14.3 for. 6/21/2012 7:58:05 AM -- Instance4956.4 has 0.2 for Log. 6/21/2012 7:58:05 AM -- Instance4956.4 has 0.2 for Log. 6/21/2012 7:58:05 AM -- Instance4956.5 has 14.0 for. 6/21/2012 7:58:05 AM -- Instance4956.5 has 0.2 for Log. 6/21/2012 7:58:05 AM -- Instance4956.5 has 0.2 for Log. 6/21/2012 7:58:05 AM -- Instance4956.6 has 14.2 for. 6/21/2012 7:58:05 AM -- Instance4956.6 has 0.2 for Log. 6/21/2012 7:58:05 AM -- Instance4956.6 has 0.2 for Log. 6/21/2012 7:58:05 AM -- Test has 0 Maximum Page Fault Stalls/sec. 6/21/2012 7:58:05 AM -- The test has 0 Page Fault Stalls/sec samples higher than 0. 6/21/2012 7:58:05 AM -- C:\Jetstress tests\2000 user 4 threads\performance_2012_6_20_14_14_52.xml has 479 samples queried. 12
24 hour test Test Summary Overall Test Result Machine Name Test Description Test Start Time Test End Time Pass D2600 24 Hour Stress Test 2000 users, 0.24 IOPS, 4GB Mailbox, TC=4 6/21/2012 3:49:15 AM 6/23/2012 7:57:59 PM Collection Start Time 6/21/2012 3:50:57 AM Collection End Time 6/22/2012 3:50:50 AM Jetstress Version 14.01.0225.017 ESE Version 14.02.0283.000 Operating System Windows Server 2008 R2 Enterprise Service Pack 1 (6.1.7601.65536) Performance Log C:\Jetstress tests\2000 user 4 threads\performance_2012_6_21_8_54_11.blg Sizing and Throughput Achieved Transactional per Second 702.833 Target Transactional per Second 480 Initial Size (bytes) 8616081227776 Final Size (bytes) 8623496757248 Files (Count) 6 Jetstress System Parameters Thread Count Minimum Cache Maximum Cache Insert Operations 40% Delete Operations 20% Replace Operations 5% Read Operations 35% Lazy Commits 70% Run Background Maintenance True Number of Copies per 2 4 (per database) 192.0 MB 1536.0 MB Configuration Instance4956.1 Log path: E:\dblg1 : E:\dblg1\Jetstress001001.edb Instance4956.2 Log path: F:\dblg2 : F:\dblg2\Jetstress002001.edb Instance4956.3 Log path: G:\dblg3 : G:\dblg3\Jetstress003001.edb Instance4956.4 Log path: H:\dblg4 : H:\dblg4\Jetstress004001.edb Instance4956.5 Log path: I:\dblg5 : I:\dblg5\Jetstress005001.edb Instance4956.6 Log path: J:\dblg6 : J:\dblg6\Jetstress006001.edb Transactional Performance MSExchange ==> Instances /sec /sec Log Log Log /sec Log /sec Log Log Instance4956.1 19.439 1.536 72.475 44.502 33979.413 35354.852 0.000 0.180 0.000 39.094 0.000 4461.371 Instance4956.2 14.572 1.412 72.369 44.395 34737.014 35348.660 0.000 0.172 0.000 39.085 0.000 4489.613 Instance4956.3 14.644 1.299 72.673 44.602 34454.474 35350.243 0.000 0.175 0.000 39.203 0.000 4517.859 Instance4956.4 14.501 1.183 72.423 44.418 34578.030 35376.733 0.000 0.173 0.000 39.069 0.000 4513.190 Instance4956.5 14.416 1.040 73.033 44.840 34561.193 35346.953 0.000 0.173 0.000 39.233 0.000 4490.878 Instance4956.6 14.326 0.713 72.582 44.521 34543.802 35352.108 0.000 0.173 0.000 39.225 0.000 4497.710 Background Maintenance Performance MSExchange ==> Instances Maintenance IO /sec Maintenance IO Instance4956.1 24.137 261850.973 Instance4956.2 26.308 261814.560 Instance4956.3 27.500 261866.496 Instance4956.4 26.220 261871.005 Instance4956.5 27.266 261830.224 Instance4956.6 26.286 261845.486 13
Log Replication Performance MSExchange ==> Instances Log /sec Log Instance4956.1 0.703 224644.930 Instance4956.2 0.707 223816.799 Instance4956.3 0.714 223953.439 Instance4956.4 0.711 225456.483 Instance4956.5 0.710 223813.479 Instance4956.6 0.711 223680.324 Total Performance MSExchange ==> Instances /sec /sec Log Log Log /sec Log /sec Log Log Instance4956.1 19.439 1.536 96.612 44.502 90909.760 35354.852 0.307 0.180 0.703 39.094 224644.930 4461.371 Instance4956.2 14.572 1.412 98.677 44.395 95277.190 35348.660 0.313 0.172 0.707 39.085 223816.799 4489.613 Instance4956.3 14.644 1.299 100.173 44.602 96885.692 35350.243 0.285 0.175 0.714 39.203 223953.439 4517.859 Instance4956.4 14.501 1.183 98.643 44.418 94993.859 35376.733 0.319 0.173 0.711 39.069 225456.483 4513.190 Instance4956.5 14.416 1.040 100.299 44.840 96343.806 35346.953 0.285 0.173 0.710 39.233 223813.479 4490.878 Instance4956.6 14.326 0.713 98.868 44.521 94976.837 35352.108 0.323 0.173 0.711 39.225 223680.324 4497.710 Host System Performance Counter Minimum Maximum % Processor Time 0.578 0.329 1.680 Available M 44631.408 44533.000 44749.000 Free System Page Table Entries 33555092.224 33555068.000 33555606.000 Transition Pages RePurposed/sec 0.000 0.000 0.000 Pool Nonpaged 184924858.701 184889344.000 185393152.000 Pool Paged 105001453.059 102158336.000 150032384.000 Page Fault Stalls/sec 0.000 0.000 0.000 Test Log6/3/2011 3:49:15 AM -- Jetstress testing begins... 6/21/2011 3:49:15 AM -- Preparing for testing... 6/21/2011 3:49:24 AM -- Attaching databases... 6/21/2011 3:49:24 AM -- Preparations for testing are complete. 6/21/2011 3:49:24 AM -- Starting transaction dispatch.. 6/21/2011 3:49:24 AM -- cache settings: (minimum: 288.0 MB, maximum: 2.2 GB) 6/21/2011 3:49:24 AM -- flush thresholds: (start: 23.0 MB, stop: 46.1 MB) 6/21/2011 3:49:34 AM -- read latency thresholds: (average: 20 msec/read, maximum: 100 msec/read). 6/21/2011 3:49:34 AM -- Log write latency thresholds: (average: 10 msec/write, maximum: 100 msec/write). 6/21/2011 3:49:43 AM -- Operation mix: Sessions 8, Inserts 40%, Deletes 20%, Replaces 5%, 35%, Lazy Commits 70%. 6/21/2011 3:49:43 AM -- Performance logging started (interval: 15000 ms). 6/21/2011 3:49:43 AM -- Attaining prerequisites: 6/21/2011 3:50:57 AM -- \MSExchange (JetstressWin)\ Cache Size, Last: 2185871000.0 (lower bound: 2174327000.0, upper bound: none) 6/22/2011 3:50:58 AM -- Performance logging has ended. 6/23/2011 7:57:48 PM -- JetInterop batch transaction stats: 1831815, 1831814, 1832035, 1832346, 1830663, 1831065, 1830144, 1827729 and 1829473. 6/23/2011 7:57:48 PM -- Dispatching transactions ends. 6/23/2011 7:57:48 PM -- Shutting down databases... 6/23/2011 7:57:59 PM -- Instance5052.1 (complete), Instance5052.2 (complete), Instance5052.3 (complete), Instance5052.4 (complete), Instance5052.5 (complete), Instance5052.6 (complete), Instance5052.7 (complete), Instance5052.8 (complete) and Instance5052.9 (complete) 6/23/2012 7:57:59 PM -- C:\Jetstress tests\2000 user 4 threads\performance_2012_6_21_8_54_11.blg has 1715 samples. 6/23/2012 7:57:59 PM -- Creating test report... 6/23/2012 7:58:38 PM -- Instance4956.1 has 19.4 for. 6/23/2012 7:58:38 PM -- Instance4956.1 has 0.2 for Log. 6/23/2012 7:58:38 PM -- Instance4956.1 has 0.2 for Log. 6/23/2012 7:58:38 PM -- Instance4956.2 has 14.6 for. 6/23/2012 7:58:38 PM -- Instance4956.2 has 0.2 for Log. 6/23/2012 7:58:38 PM -- Instance4956.2 has 0.2 for Log. 6/23/2012 7:58:38 PM -- Instance4956.3 has 14.6 for. 6/23/2012 7:58:38 PM -- Instance4956.3 has 0.2 for Log. 6/23/2012 7:58:38 PM -- Instance4956.3 has 0.2 for Log. 6/23/2012 7:58:38 PM -- Instance4956.4 has 14.5 for. 6/23/2012 7:58:38 PM -- Instance4956.4 has 0.2 for Log. 6/23/2012 7:58:38 PM -- Instance4956.4 has 0.2 for Log. 6/23/2012 7:58:38 PM -- Instance4956.5 has 14.4 for. 6/23/2012 7:58:38 PM -- Instance4956.5 has 0.2 for Log. 6/23/2012 7:58:38 PM -- Instance4956.5 has 0.2 for Log. 6/23/2012 7:58:38 PM -- Instance4956.6 has 14.3 for. 6/23/2012 7:58:38 PM -- Instance4956.6 has 0.2 for Log. 6/23/2012 7:58:38 PM -- Instance4956.6 has 0.2 for Log. 6/23/2012 7:58:38 PM -- Test has 0 Maximum Page Fault Stalls/sec. 6/23/2012 7:58:38 PM -- The test has 0 Page Fault Stalls/sec samples higher than 0. 6/23/2012 7:58:38 PM -- C:\Jetstress tests\2000 user 4 threads \Performance_2012_6_21_8_54_11.xml has 1702 samples queried. 14
backup test Backup Statistics - All Instance Size (M) Elapsed Backup Time M Transferred/sec Instance4956.1 1370642.59 11:03:37 34.42 Instance4956.2 1370658.59 09:38:25 39.49 Instance4956.3 1370674.59 09:52:23 38.56 Instance4956.4 1370650.59 10:09:27 37.48 Instance4956.5 1370674.59 09:56:35 38.29 Instance4956.6 1370658.59 09:56:08 38.32 Jetstress System Parameters Thread Count 4 (per database) Minimum Cache 192.0 MB Maximum Cache 1536.0 MB Insert Operations 40% Delete Operations 20% Replace Operations 5% Read Operations 35% Lazy Commits 70% Configuration Instance4956.1 Log path: E:\dblg1 Backup path: E:\dblg1 : E:\dblg1\Jetstress001001.edb Instance4956.2 Log path: F:\dblg2 Backup path: F:\dblg2 : F:\dblg2\Jetstress002001.edb Instance4956.3 Log path: G:\dblg3 Backup path: G:\dblg3 : G:\dblg3\Jetstress003001.edb Instance4956.4 Log path: H:\dblg4 Backup path: H:\dblg4 : H:\dblg4\Jetstress004001.edb Instance4956.5 Log path: I:\dblg5 Backup path: I:\dblg5 : I:\dblg5\Jetstress005001.edb Instance4956.6 Log path: J:\dblg6 Backup path: J:\dblg6 : J:\dblg6\Jetstress006001.edb Transactional Performance MSExchange ==> Instances /sec /sec Log Log Log /sec Log /sec Log Log Instance4956.1 14.590 0.000 137.645 0.000 262144.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Instance4956.2 12.239 0.000 157.881 0.000 262144.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Instance4956.3 12.443 0.000 154.135 0.000 262144.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Instance4956.4 13.037 0.000 149.737 0.000 262144.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Instance4956.5 12.605 0.000 153.122 0.000 262144.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Instance4956.6 12.462 0.000 153.171 0.000 262144.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 Host System Performance Counter Minimum Maximum % Processor Time 1.647 0.138 9.719 Available M 36556.803 28410.000 44561.000 Free System Page Table Entries 33555423.659 33555089.000 33556630.000 Transition Pages RePurposed/sec 52667.285 0.000 76758.493 Pool Nonpaged 291893529.098 184979456.000 370069504.000 Pool Paged 8212032047.107 126050304.000 16239865856.000 Page Fault Stalls/sec 0.000 0.000 0.000 15
Test Log6/22/2012 8:13:15 AM -- Jetstress testing begins... 6/22/2012 8:13:15 AM -- Preparing for testing... 6/22/2012 8:13:22 AM -- Attaching databases... 6/22/2012 8:13:22 AM -- Preparations for testing are complete. 6/22/2012 8:13:31 AM -- Performance logging started (interval: 30000 ms). 6/22/2012 8:13:31 AM -- Backing up databases... 6/22/2012 7:17:08 PM -- Performance logging has ended. 6/22/2012 7:17:08 PM -- Instance4956.1 (100% processed), Instance4956.2 (100% processed), Instance4956.3 (100% processed), Instance4956.4 (100% processed), Instance4956.5 (100% processed) and Instance4956.6 (100% processed) 6/22/2012 7:17:08 PM -- C:\Jetstress tests\db Backup test\backup_2012_6_22_8_13_22.blg has 1326 samples. 6/22/2012 7:17:08 PM -- Creating test report... Soft recovery test Soft-Recovery Statistics - All Instance Log files replayed Elapsed seconds Instance3728.1 507 1925.8722846 Instance3728.2 501 1698.8606859 Instance3728.3 505 1837.5605295 Instance3728.4 501 1795.6588559 Instance3728.5 507 1825.6265085 Instance3728.6 502 1824.8309071 Configuration Instance3728.1 Log path: E:\dblg1 : E:\dblg1\Jetstress001001.edb Instance3728.2 Log path: F:\dblg2 : F:\dblg2\Jetstress002001.edb Instance3728.3 Log path: G:\dblg3 : G:\dblg3\Jetstress003001.edb Instance3728.4 Log path: H:\dblg4 : H:\dblg4\Jetstress004001.edb Instance3728.5 Log path: I:\dblg5 : I:\dblg5\Jetstress005001.edb Instance3728.6 Log path: J:\dblg6 : J:\dblg6\Jetstress006001.edb Transactional Performance MSExchange ==> Instances /sec /sec Log Log Log /sec Log /sec Log Instance3728.1 31.898 0.146 201.504 1.565 39712.223 30784.134 12.585 0.000 2.348 0.000 216293.968 0.000 Instance3728.2 26.897 0.152 230.403 1.752 39495.015 31991.507 14.238 0.000 2.628 0.000 225673.565 0.000 Instance3728.3 28.938 0.150 211.628 1.634 39476.706 31549.059 15.325 0.000 2.451 0.000 223249.550 0.000 Instance3728.4 27.884 0.136 217.222 1.654 39500.766 31445.525 11.688 0.000 2.481 0.000 223225.070 0.000 Instance3728.5 28.063 0.591 211.880 1.649 39490.763 31252.300 12.564 0.000 2.472 0.000 221847.152 0.000 Instance3728.6 27.756 0.586 215.655 1.645 39533.731 31324.476 12.081 0.000 2.467 0.000 221930.935 0.000 Background Maintenance Performance MSExchange ==> Instances Maintenance IO /sec Maintenance IO Instance3728.1 23.789 261964.328 Instance3728.2 22.999 261901.856 Instance3728.3 24.784 261829.348 Instance3728.4 22.084 261840.030 Instance3728.5 24.450 261919.446 Instance3728.6 22.311 261922.509 Log 16
Total Performance MSExchange ==> Instances /sec /sec Log Log Log /sec Log /sec Log Instance3728.1 31.898 0.146 225.293 1.565 63180.448 30784.134 12.585 0.000 2.348 0.000 216293.968 0.000 Instance3728.2 26.897 0.152 253.402 1.752 59681.089 31991.507 14.238 0.000 2.628 0.000 225673.565 0.000 Instance3728.3 28.938 0.150 236.413 1.634 62787.141 31549.059 15.325 0.000 2.451 0.000 223249.550 0.000 Instance3728.4 27.884 0.136 239.306 1.654 60019.331 31445.525 11.688 0.000 2.481 0.000 223225.070 0.000 Instance3728.5 28.063 0.591 236.330 1.649 62502.438 31252.300 12.564 0.000 2.472 0.000 221847.152 0.000 Instance3728.6 27.756 0.586 237.966 1.645 60383.850 31324.476 12.081 0.000 2.467 0.000 221930.935 0.000 Host System Performance Counter Minimum Maximum % Processor Time 0.689 0.000 2.010 Available M 42658.475 42567.000 43968.000 Free System Page Table Entries 33555604.196 33555603.000 33555606.000 Transition Pages RePurposed/sec 17.074 0.000 4535.049 Pool Nonpaged 375487368.533 375484416.000 375500800.000 Pool Paged 2973411208.533 2973392896.000 2973581312.000 Page Fault Stalls/sec 0.000 0.000 0.000 Log Test Log6/28/2012 8:31:23 AM -- Jetstress testing begins... 6/28/2012 8:31:23 AM -- Preparing for testing... 6/28/2012 8:31:29 AM -- Attaching databases... 6/28/2012 8:31:29 AM -- Preparations for testing are complete. 6/28/2012 8:31:29 AM -- Starting transaction dispatch... 6/28/2012 8:31:29 AM -- cache settings: (minimum: 192.0 MB, maximum: 1.5 GB) 6/28/2012 8:31:29 AM -- flush thresholds: (start: 15.3 MB, stop: 30.7 MB) 6/28/2012 8:31:36 AM -- read latency thresholds: (average: 20 msec/read, maximum: 100 msec/read). 6/28/2012 8:31:36 AM -- Log write latency thresholds: (average: 10 msec/write, maximum: 100 msec/write). 6/28/2012 8:31:39 AM -- Operation mix: Sessions 4, Inserts 40%, Deletes 20%, Replaces 5%, 35%, Lazy Commits 70%. 6/28/2012 8:31:39 AM -- Performance logging started (interval: 15000 ms). 6/28/2012 8:31:39 AM -- Generating log files... 6/28/2012 10:15:47 AM -- E:\dblg1 (101.4% generated), F:\dblg2 (100.2% generated), G:\dblg3 (101.0% generated), H:\dblg4 (100.2% generated), I:\dblg5 (101.4% generated) and J:\dblg6 (100.4% generated) 6/28/2012 10:15:47 AM -- Performance logging has ended. 6/28/2012 10:15:47 AM -- JetInterop batch transaction stats: 21754, 21798, 21831, 21602, 21618 and 21601. 6/28/2012 10:15:47 AM -- Dispatching transactions ends. 6/28/2012 10:15:47 AM -- Shutting down databases... 6/28/2012 10:15:48 AM -- Instance3728.1 (complete), Instance3728.2 (complete), Instance3728.3 (complete), Instance3728.4 (complete), Instance3728.5 (complete) and Instance3728.6 (complete) 6/28/2012 10:15:48 AM -- C:\Program Files\Exchange Jetstress\Performance_2012_6_28_8_31_36.blg has 416 samples. 6/28/2012 10:15:48 AM -- Creating test report... 6/28/2012 10:15:49 AM -- Instance3728.1 has 18.2 for. 6/28/2012 10:15:49 AM -- Instance3728.1 has 0.2 for Log. 6/28/2012 10:15:49 AM -- Instance3728.1 has 0.2 for Log. 6/28/2012 10:15:49 AM -- Instance3728.2 has 15.6 for. 6/28/2012 10:15:49 AM -- Instance3728.2 has 0.2 for Log. 6/28/2012 10:15:49 AM -- Instance3728.2 has 0.2 for Log. 6/28/2012 10:15:49 AM -- Instance3728.3 has 16.4 for. 6/28/2012 10:15:49 AM -- Instance3728.3 has 0.2 for Log. 6/28/2012 10:15:49 AM -- Instance3728.3 has 0.2 for Log. 6/28/2012 10:15:49 AM -- Instance3728.4 has 16.5 for. 6/28/2012 10:15:49 AM -- Instance3728.4 has 0.2 for Log. 6/28/2012 10:15:49 AM -- Instance3728.4 has 0.2 for Log. 6/28/2012 10:15:49 AM -- Instance3728.5 has 16.4 for. 6/28/2012 10:15:49 AM -- Instance3728.5 has 0.2 for Log. 6/28/2012 10:15:49 AM -- Instance3728.5 has 0.2 for Log. 6/28/2012 10:15:49 AM -- Instance3728.6 has 16.3 for. 6/28/2012 10:15:49 AM -- Instance3728.6 has 0.2 for Log. 6/28/2012 10:15:49 AM -- Instance3728.6 has 0.2 for Log. 6/28/2012 10:15:49 AM -- Test has 0 Maximum Page Fault Stalls/sec. 6/28/2012 10:15:49 AM -- The test has 0 Page Fault Stalls/sec samples higher than 0. 6/28/2012 10:15:49 AM -- C:\Program Files\Exchange Jetstress\Performance_2012_6_28_8_31_36.xml has 415 samples queried. 6/28/2012 10:15:49 AM -- C:\Program Files\Exchange Jetstress\Performance_2012_6_28_8_31_36.html was saved. 6/28/2012 11:16:02 AM -- Performance logging started (interval: 4000 ms). 6/28/2012 11:16:02 AM -- Recovering databases... 6/28/2012 11:48:08 AM -- Performance logging has ended. 6/28/2012 11:48:08 AM -- Instance3728.1 (1925.8722846), Instance3728.2 (1698.8606859), Instance3728.3 (1837.5605295), Instance3728.4 (1795.6588559), Instance3728.5 (1825.6265085) and Instance3728.6 (1824.8309071) 6/28/2012 11:48:08 AM -- C:\Program Files\Exchange Jetstress\SoftRecovery_2012_6_28_11_15_59.blg has 480 samples. 6/28/2012 11:48:08 AM -- Creating test report... 17
For more information QuickSpecs for HP D2000 disk enclosures: http://h18000.www1.hp.com/products/quickspecs/13404_div/13404_div.html QuickSpecs for HP Smart Array P411 controller: http://h18000.www1.hp.com/products/quickspecs/13202_na/13202_na.html To read more about HP solutions for Exchange, including best practices and additional testing of HP server and storage products with Exchange, see: hp.com/solutions/exchange For general information on Exchange sizing and best practices, there are several white papers and planning tools that are free to download from: hp.com/solutions/activeanswers/exchange For more information on Jetstress and the process and requirements for validating an Exchange storage solution, see: http://gallery.technet.microsoft.com/jetstress-field-guide-1602d64c To help us improve our documents, please provide feedback at hp.com/solutions/feedback. Get connected hp.com/go/getconnected Current HP driver, support, and security alerts delivered directly to your desktop Copyright 2012 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 omiss ions contained herein. Microsoft and Windows are U.S. registered trademarks of Microsoft Corporation. 18 4AA4-3116ENW, Created September 2012; Updated September 2012, Rev. 1