XADM: Overview of EMC TimeFinder/Exchange Backup/Recovery Scripts

ID: Q236347

The information in this article applies to:
  • Microsoft Exchange Server, versions 5.5, 5.5 SP2

SUMMARY

This article briefly documents the steps executed during backup and recovery, using EMC TimeFinder scripts with Microsoft Exchange Server 5.5.


MORE INFORMATION

This specific implementation of Microsoft Exchange Server 5.5 and EMC TimeFinder 3.2, including the procedures and statements contained in this document, have been reviewed by Microsoft and EMC. Combined use of Microsoft Exchange Server 5.5 and EMC TimeFinder software, in accordance with these procedures, satisfies the requirements for a supported environment. EMC provides first-level support for service calls related to implementation and use of procedures outlined within.

Basic Configuration of EMC TimeFinder and Exchange Server

EMC TimeFinder works by creating multiple independently-addressable online Business Continuance Volumes (BCVs) for information storage. The BCV is configured as a single mirror and independently supports host applications and processes (for example, backups). Microsoft Exchange Server will be installed on the EMC Symmetrix, on a production volume, with RAID 1 mirroring for the Exchange Server components.

NOTE: Exchange Server databases and logs should be installed on separate disks on the production volume. Additionally, there will be a backup and recovery server that will use the BCVs to perform backups to tape and other operations.

Backup Process

The following steps outline the automated process used during a backup to a BCV. This process is automated in a set of PERL scripts created by EMC.
  1. The BCV volume to synchronize with the production volumes is established. During this process, only tracks that have changed since the previous 'establish' are copied.


  2. The Microsoft Exchange Server services are stopped. In addition, a command is executed to ensure the entire Windows NT host cache is flushed to disk. Stopping the Exchange Server services successfully ensures the database is in a consistent state when synchronized to the BCV volume.


  3. The BCV backup set volumes are split from the production server volumes. Now that the consistent databases and all log files are backed up to the BCV volume, the BCV volume is split off to ensure no other changes.


  4. The Microsoft Exchange Server services are restarted.


  5. The BCV copy is verified. During this step, Esefile (which is included with the EMC TimeFinder scripts) is run against the Priv.edb, Pub.edb, and Dir.edb databases to verify the physical consistency of the database.

    NOTE: Use the backup and recovery server to perform this process by mounting the BCV to the backup and recovery server.

    The Esefile.exe is available on the Microsoft Exchange Server 5.5 Service Pack 3 CD, or may be obtained from Microsoft Product Support Services. Use Esefile to verify the page-level integrity of a copied Exchange Server 5.0 or 5.5 database.


  6. The last consistent transaction log is determined. Run eseutil /MH to dump the database header. The last consistent log generation is then stored for all Exchange Server databases for later reference.


  7. The BCV backup set is copied to tape.

    NOTE: The backup and recovery server processor is used for backup, not on the production server processor.


  8. Using the information from last consistent log generation (step 6), unneeded transaction logs are purged from the production log volume after the backup is successfully completed.

    NOTE: The Edb.log file is never purged because it holds the current generation number for the log files.


  9. Remove the Exchange Server database and transaction log volumes from the backup and recovery server.


Restore Process

The following steps outline the automated process used during a restore from a BCV backup set. This process is automated in a set of PERL scripts created by EMC.
  1. The production volumes are restored from the BCV copy. During this process, only tracks that have changed since the last split are copied from the BCV volume to the production volume. After the data is restored to the production server, the BCV volume is split from the production volume.


  2. Information Store Transaction Log Verification. Prior to restarting the Exchange Server services and initiating log recovery, the scripts perform a simple log verification. The verification ensures all the necessary .log files are available to complete recovery. Then, it ensures the .log files are sequential. Finally, it verifies that the database signature in the .log files is consistent with the signature in the database header. The .log file verification is accomplished using Eseutil to dump the headers of the files.


  3. Directory Transaction Log Verification. The same verification process is performed on the directory .log files.


  4. Delete the checkpoint file. Because the recovered databases are "consistent," you need to delete the Edb.chk file. Then, when you start the Exchange Server services, they begin to replay the .log files.


The third-party products discussed here are manufactured by vendors independent of Microsoft; we make no warranty, implied or otherwise, regarding these products' performance or reliability.

Additional query words: Sym Time Finder corruption Symm Symmetrix

Keywords : exc55 exc55sp2
Version : winnt:5.5,5.5 SP2
Platform : winnt
Issue type : kbhowto


Last Reviewed: September 8, 1999
© 2000 Microsoft Corporation. All rights reserved. Terms of Use.