Protect Microsoft Exchange 2013 with Hitachi Application Protector in a large enterprise environment

Document created by leo nguyen on Aug 14, 2014
Version 1Show Document
  • View in full screen mode


This tech note shows how you can use Hitachi Application Protector on Hitachi Unified Storage VM to protect your Microsoft® Exchange databases from logical corruption or accidental deletion in a large enterprise environment.

The Exchange environment used to validate this solution consisted of the following:

§  Four-member Exchange database availability group (DAG)

§  24,000 × 500 MB size mailboxes

§  216 × 550 GB LUNs for Exchange databases and logs

§  Three database copies, one active and two passive

§  Seven day snapshot rotation counts

Application Protector performs rapid backups and restores of Exchange databases on a standalone server or on a database availability group environment. It uses Hitachi Thin Image snapshot technology and Volume Shadow Copy Service from Microsoft to give application-consistent data protection. Application Protector provides the following advantages for Exchange:

§  Provides application consistency

§  Protects active or passive database copy at the snapshot or full-copy level

§  Supports Exchange database availability groups

§  Eliminates application strain during backups using Volume Shadow Copy Service

§  Reduces backup window and improves recovery point objective and recovery time objective

§  Uses Kroll Ontrack PowerControls for granular recovery

NoteTesting of this configuration was in a lab environment. Many things affect production environments beyond prediction or duplication in a lab environment. Follow the recommended practice of conducting proof-of-concept testing for acceptable results in a non-production, isolated test environment that otherwise matches your production environment before your production implementation of this solution.

Solution Components

Hitachi Application Protector uses Hitachi Thin Image snapshot technology and Volume Shadow Copy Service from Microsoft to give you application-consistent data protection for Microsoft Exchange 2013. Application Protector recovers databases using snapshots based on recovery time objective or the recovery point objective.

Figure 1 shows the physical architecture for protecting Exchange 2013 in a DAG environment using Hitachi Application Protector.

physical architecture.png

                           

Figure 1

 

Table 1 lists the hardware components used in this solution.

Table 1. Hardware Components

           

Hardware

Description

Version

Quantity

Hitachi Unified Storage VM

§  Dual controllers

§    16 × 8 Gb/sec Fibre Channel ports

§  128 GB cache memory

§  200 × 3TB SAS disks

73-03-01-00

1

Hitachi Compute Blade 500 chassis

§  8-blade chassis

§  2 management modules

§  6 cooling fan modules

§  4 power supply modules

§  2 Brocade 5460 Fibre Channel switch  modules

§  2 Brocade 10GbE DCB switch modules

SVP: A0135-D-6829

1

520HB2 Server Blade

§  Half blade

§    2 × 12-core Intel Xeon E5-2697 v2  processor, 2.70 GHz

§  384 GB RAM

§  Emulex 10GbE CNA onboard network  adapter

§    Hitachi 8Gb/sec Fibre Channel Mezzanine  Card

04-05

4

 

Table 2 lists the software components used in this solution.

Table 2. Software Components

           

Software

Version

Hitachi Application Protector

2.8.0.41

Command Control Interface

01-31-03-07

Hitachi Storage Adapter for  Microsoft® Volume ShadowCopy Service

4.11

Microsoft Windows Server®

2012 R2 Standard

Microsoft Exchange Server

2013 SP1 Enterprise

Test Environment Configuration

The test environment consisted of the following hardware and software configurations.

Hitachi Unified Storage VM Configuration

200 × 3 TB SAS 7.2k rpm disks were configured as RAID-10 (2D+2D). Two Hitachi Dynamic Provisioning pools were configured for active and passive databases. One Thin Image pool was configured for snapshots. One 1 GB command device was configured and presented to all four test hosts. The storage configuration details for the pools are listed in Table 3.

Table 3. Hitachi Unified Storage VM Storage Configuration

       

HDP  Pool

Number  of Drives

Pool  Capacity (TB)

Exchange active database

40

53.42

Exchange passive database

80

106.85

Thin Image

80

106.85

Exchange Configuration

A four-node Exchange 2013 DAG was configured to support 24,000 500 MB mailboxes using three database copies for database resiliency. Each server had 18 active databases and 36 passive databases. Each database had 500 mailbox users. The databases and logs were configured on separate LUNs. Each database and log LUN were 550 GB. The database distribution details for Exchange are listed in Table 4.

Table 4. Exchange Database Distribution

               

Active/Passive DB

Server 1

Server 2

Server 3

Server 4

Active database

Db1-18

Db19-36

Db37-54

Db55-72

Passive copy 1

Db55-72

Db1-18

Db19-36

Db37-54

Passive copy 2

Db37-54

Db55-72

Db1-18

Db19-36

Host Configuration

Command control interface (CCI), Hitachi Storage Adapter for Microsoft® Volume ShadowCopy Service and Hitachi Application Protector (HAPRO) server for Exchange were installed and configured on each host. HAPRO was configured to run a full snap backup of all the active databases every six hours and all the passive databases every eight hours. The schedule was to rotate and delete old snapshots after seven counts.

Test Cases Tested

These are the use cases tested for protecting Microsoft Exchange 2013 with Hitachi Application Protector.

Schedule recurring full snapshot backup of all 18 active databases on server 1.

The objective of this test was to schedule recurring full snapshot backups of all 18 active databases.

All eight sets of backups completed successfully. They took an average of 20 minutes.

Schedule recurring full snapshot backup of all 36 passive databases on server 1.

The objective of this test was to schedule recurring full snapshot backups of all 36 passive databases.

All eight sets of backups completed successfully. They took an average of 55 minutes.

Schedule recurring full snapshot backup of all 18 active databases on server 2.

The objective of this test was to schedule recurring full snapshot backups of all 18 active databases.

All eight sets of backups completed successfully. They took an average of 23 minutes.

Schedule recurring full snapshot backup of all 36 passive databases on server 2.

The objective of this test was to schedule recurring full snapshot backups of all 36 passive databases.

All eight sets of backups completed successfully. They took an average of 50 minutes.

Schedule recurring full snapshot backup of all 18 active databases on server 3.

The objective of this test was to schedule recurring full snapshot backups of all 18 active databases.

All eight sets of backups completed successfully. They took an average of 14 minutes.

Schedule recurring full snapshot backup of all 36 passive databases on server 3.

The objective of this test was to schedule recurring full snapshot backups of all 36 passive databases.

All eight sets of backups completed successfully. They took an average of 30 minutes.

Schedule recurring full snapshot backup of all 18 active databases on server 4.

The objective of this test was to schedule recurring full snapshot backups of all 18 active databases.

All eight sets of backups completed successfully. They took an average of 17 minutes.

Schedule recurring full snapshot backup of all 36 passive databases on server 4.

The objective of this test was to schedule recurring full snapshot backups of all 36 passive databases.

All eight sets of backups completed successfully. They took an average of 36 minutes.

Mount and browse a snapshot for 18 active databases.

The objective of this test was to mount and browse a snapshot for all the active databases to the default directory C:\HAPROMount.

The test completed successfully. It took 2 minutes to mount a snapshot.

Unmount snapshot.

The objective of this test was to unmount a snapshot.

The test completed successfully. It took 3 minutes to unmount a snapshots.

Mount and browse a snapshot for 36 passive databases.

The objective of this test was to mount and browse a snapshot for all the passive databases to the default directory C:\HAPROMount.

The test completed successfully. It took 4 minutes to mount the snapshots.

Unmount a snapshot.

The objective of this test was to unmount a snapshot.

The test completed successfully. It took 5 minutes to unmount the snapshots.

Recover by overwriting the existing database.

The objective of this test was to recover db55 from the snap backup and overwrite the existing database. Usually this scenario is used when the database is corrupted and can't be mounted. In HAPRO restore wizard, check Enable Exchange Server Overwrite Flag and The target MailStore database will be overwritten with the contents of the MailStore to the recovered. 

Restore can only be done to active databases. Restore to passive databases is not supported.

The test completed successfully. It took 52 minutes to restore db55.

Recover by preserving existing database.

The objective of this test was to restore db55 by preserving the existing database. The original database and log files were renamed. Ensure there is sufficient disk space on the destination volume.

The test completed successfully. It took 51 minutes to restore db55.

Delete a snapshot.

The objective of this test was to delete a snapshot containing passive db37-72.

The test completed successfully. It took 3 hours and 32 minutes to delete a snapshot.

Attachments

    Outcomes