SEPsesam for Microsoft Exchange Server 2007
(C)SEP AG
Copyright 1999-2011 by SEP AG. All Rights reserved.
Any form of reproduction of the contents or parts of this manual is allowed only with the express written permission from SEP AG. When compiling and designing user documentation SEP AG uses great diligence and attempts to deliver accurate and correct information. However, SEP AG cannot issue a guarantee for the contents of this manual.
SEPsesam online module for Exchange Server 2007 provides an online backup of Exchange Server 2007. This includes saving and restoring Mailbox databases.
This document describes the necessary steps to back up the different Exchange 2007 Server roles and how Exchange Server 2007 restore and recovery may be done.
Note It is possible to install and use SEP sesam Server together with the Exchange Server on one system. However SEP recommends to use a separate system because otherwise a Disaster Recovery may become very complex and time consuming.
The most important data to back up are the databases, which store user mailboxes and public data.
Exchange Server 2007 introduced a new continuous replication functionality to keep a copy of one or more databases in a storage group in sync with the active versions of the databases using log file shipping and replay. This new feature is provided as Local continuous replication (LCR) for a local copy, Cluster continuous replication (CCR) to allow a copy to a Microsoft Cluster node or Standby continuous replication (SCR) to copy to a standby node (See article "High Availability" under http://technet.microsoft.com/en-us/library/bb124721.aspx ). Although the new feature provides an additional protection for Mailbox and Public Folder databases it does not substitute a database backup to recover from in case of system disaster. Moreover the log files which are generated by the Exchange Server are kept on disks until a backup will commit and delete them. If log files aren't deleted they may occupy all disk space and then the database will be dismounted.
Note A target storage group in a standby continuous replication (SCR) environment cannot be used to backup Exchange 2007 Mailbox Server. Only a backup on the passive CCR cluster node initiates log truncation by communicating with Microsoft Exchange IS Service on the active node.
To allow differential and incremental backups Circular Logging must be disabled for every Storage Group. See article "How to Enable or Disable Circular Logging for a Storage Group" http://technet.microsoft.com/en-us/library/bb331968.aspx.
Since an Exchange 2007 Mailbox Server can always be recovered using the Setup /Mode:RecoverServer command the System State backup for the Exchange Mailbox Server can be omitted. SEP recommends to do a regular system_state backup nevertheless.
Exchange 2007 supports two different backup methods to backup databses. The first is the streaming backup method based on the ESE application programming interface (ESE API), which was already provided in previous versions of Exchange. ESE API allows to back up one or more storage groups at the same time, with the restriction that only one backup job can run against a storage group. In Exchange Server 2003 a new backup method via Volume Shadow Copy Service (VSS) was implemented with the "Exchange VSS Writer". This new feature allows to backup the Exchange Server by generating a volume snapshot. The backup application must be VSS-aware and must handle this snapshot.
The ESE API is only available on the active node however VSS is available on both nodes (depending on "Exchange VSS Writer" or on the passive copy node "Exchange Replica VSS Writer").
| Active Node | Passive Node |
| ESE API | - |
| VSS | VSS |
SEP sesam Client for Exchange Server is using streaming backup method via ESE API.
Starting with SEP sesam version 3.4.1.88, SBC Version 3.2A14, SEP sesam Backup client program sbc.exe can create and save VSS snapshots (depending on the switch '-o vss'). These snapshots may be used for a system disaster recovery. If the "Exchange VSS Writer" or "Exchange Replica VSS Writer" is working properly the backup does allow to restore the Exchange Server at the point-in-time of saveset creation.
As VSS becomes more powerful since the first version in Exchange Server 2003 SEP intends to implement also the VSS backup method in a future Sesam Client for Exchange Server version.
(See article "Database Backup and Restore" http://technet.microsoft.com/en-us/library/bb124515.aspx ).
(See article "Troubleshooting Exchange 2007 VSS Backups" http://msexchangeteam.com/archive/2008/08/25/449684.aspx).
After a successful configuration of the Exchange Server as a Sesam client a Backup Task must be created.
Step-by-step Configuration:
Example backup task source to back up database:
To execute the task it should be assigned to a cyclic backup event.
The Following Screenshot shows an Complete Backup of the Exchange Environment by only specifying the Exchange Server Name.
After a Full or Incr Backup the ESE function HrESEBackupTruncateLogs is called. The HrESEBackupTruncateLogs function examines the storage group log files, and deletes those that are backed up and no longer needed to fully restore the storage group. In case of Diff or Copy Backup the log files are not deleted. Thus multiple backups of the log files are possible.
Since the Hub Transport Server stores all configuration data in the Active Directory (AD) configuration container a regular backup of the System State (Sesam backup type 'Path' with source 'system_state') must be established for the Exchange Server and the AD Server.
The AD Server must be configured as a Sesam client and a 'system_state' backup for AD Server and Exchange Server must be added. Furthermore the backup tasks must be added to a cyclic backup event.
Note Since an Exchange 2007 Hub Transport Server can always be recovered using the Setup /Mode:RecoverServer command the System State backup for the Exchange Hub Transport Server can be omitted. Nevertheless SEP recommends to do regularly system_state backups.
The Message Queues which are stored in an ESE database on an Exchange 2007 Hub Transport Server must not be backed up. There is no need for it to be backed up since message queues can be mounted on another existing, or newly installed, Hub Transport Server. Only the file mail.que, which, by default, is located under "C:\Program Files\Microsoft\Exchange Server\TransportRoles\data\Queue" must be transferred from the failed Hub Transport server to the new one.
Note Instructions for message queue handling are given under “Working with the Queue Database on Transport Servers” in the Exchange 2007 Documentation Help file.
The Message Tracking and Protocol logs of the Exchange 2007 Hub Transport Server are located under C:\Program Files\Microsoft\Exchange Server\TransportRoles\Logs. To back up these files a Sesam Path backup task can be configured.
An Exchange 2007 Edge Transport Server needs a Cloned Configuration for restore. See "Using Cloned Configuration Tasks for Edge Transport Server Disaster Recovery" http://technet.microsoft.com/en-us/library/bb125150.aspx.
The mentioned script ExportEdgeConfig.ps1 can be executed with Sesam Backup PREPOST Interface before backing up the created data.
Like on Hub Transport server the Message Queues are stored in an ESE database. They can be easily mounted on another Edge Transport server. Only the file mail.que, which, by default, is located under "C:\Program Files\Microsoft\Exchange Server\TransportRoles\data\Queue" must be transferred from the failed Hub Transport server to the new one.
In addition a System_State Backup and the Message Tracking and protocol logs under C:\Program Files\Microsoft\Exchange Server\TransportRoles\Logs should be backed up.
In the article "How to Back Up and Recover a Client Access Server" http://technet.microsoft.com/en-us/library/bb124359.aspx the full procedure is described.
To back up the Internet Information Services (IIS) metabase of a Client Access server the given 'Exchange Management Shell' command
can be executed in the Sesam PREPOST Interface before backing up the created owa.xml file.
The mentioned restore script 'restorevdir.ps1' should be prepared and tested in advance.
In addition to the IIS configuration the system_state and the following files should be backed up.
The following components are saved within the IIS metabase.
Note Like an Exchange 2007 Mailbox or Hub Transport Server a Client Access Server can be restored using the Setup /Mode:RecoverServer command.
Exchange 2007 Unified Messaging (UM) Server stores most of the configuration data in the Active Directory. Beside this several files should be backed up:
The System State should be backed up.
As the configuration data is stored in Active Directory it is possible to restore using the Setup /Mode:RecoverServer command.
If database files are corrupt or some other disaster occurs a restore may be the last opportunity to bring the system back into a productive state.
There are many different situations when a restore may be necessary. They are described under "Disaster Recovery" http://technet.microsoft.com/en-us/library/aa998848.aspx.
The required steps to prepare an Exchange Server for restore for the different scenarios are described under "Disaster Recovery Procedures" http://technet.microsoft.com/en-us/library/aa998890.aspx.
Note: Before starting a restore to an productive database the alternative use of a "Recovery Storage Groups" (RSG) should be checked. See http://technet.microsoft.com/en-us/library/aa997260.aspx. The restore procedure on Sesam side is unaffected.
To enable the restore of an Exchange Server database the following two steps must be done in 'Exchange Management Console':
The following images are showing the database dismount operation and the property "This database can be overwritten by a restore".
To restore the data of an Exchange backup the Sesam Restore Wizard GUI must be launched.
There are two main possibilities to restore the Exchange Server databases:
There are 3 recovery modes:
To simplify the restore with recovery in most cases 'With Recovery and Online' will be the best choice.
The following image shows the selection of one Exchange Server 2007 Mailbox database together with database header file, database file (*.edb) and log files.
Hint If you are not sure which header belongs to a Mailbox database then you can select all header files. Exchange Server 2007 skips all database data belonging to databases which are mounted and/or property "This database can be overwritten by a restore" is not set.
Note After a successful restore done without the recovery mode 'With Recovery and Online' the database of Exchange has to be mounted in 'Exchange Management Console'. If Exchange Server log files are available and Recovery was not yet done they will be applied automatically by the Exchange Server.
Note In case of restore from differential and incremental backups the generation restore is used to avoid inconsistent Exchange databases. After last restore a recover may be done depending on recovery mode.
Note If 'eseutil' should be used to recover then the path to the 'restore.env' file must be given. The Sesam restore automatically generates the 'restore.env' file under the installation path of the Exchange Server. Default path is "C:\Program Files\Microsoft\Exchange Server\Mailbox\First Storage Group\restore".
Note If Exchange Server does not restore the required database then check if database is dismounted and check box "This database can be overwritten by a restore" is set on the target Exchange Server or in "Recovery Storage Groups" RSG. In case of a selective restore the appropriate header, database (.edb) and log files must be restored.
Note A database can be restored and recovered to its last good state by using a full backup, differential and incremental (log files). See "Eseutil /C Restore Mode" http://technet.microsoft.com/en-us/library/aa997899.aspx.
In order to recover a corrupted database the database has to be dismounted in 'Exchange Management Console' and the check box "This database can be overwritten by a restore" must be set.
Delete all files within the directory of the database ( but not the directory itself ). Now the database is ready for restore. After the successful restore, the database must be mounted within 'Exchange Management Console' if not done with the appropriate recovery mode.
The same procedure like for "Recovery of a Corrupted Exchange Database" must be used to restore a Mailbox database to a previous state. The database has to be dismounted in 'Exchange Management Console' and the check box "This database can be overwritten by a restore" must be set. Delete all files within the directory of the database ( but not the directory itself ) and restore the Mailbox database with all required log files.
Note SEP sesam recommends to use a Recovery Storage Group (RSG) or a second Exchange Server instance to restore a Mailbox database to a previous state otherwise all recent data is removed from the Mailbox!
In the MSDN article "Using Backup to Back Up and Restore Exchange Data" http://technet.microsoft.com/en-us/library/aa998870.aspx the following description is given:
"After restoring a database from backup into a storage group that is enabled for LCR or SCR, or any storage group in a CCR environment, you must suspend and then resume continuous replication for the storage group using
and
, respectively.
This process is needed to update the Microsoft Exchange Replication Service with the correct log generation information. If continuous replication is not suspended and resumed, the Microsoft Exchange Replication Service will have outdated log generation information and will stop replicating log files."
Under "Reseeding an Exchange CCR Node" http://technet.microsoft.com/en-us/library/dd637101.aspx the full procedure including moving the database, .log and .chk files to their respective folders is described.
If the "Update Storage Group Copy" wizard is used to seed the passive node then the user will be prompted to delete any obsolete checkpoint files that are found in the passive copy of the storage group, and any existing database file that is found in the passive copy of the storage group. See "How to Seed a Cluster Continuous Replication Copy" under http://technet.microsoft.com/en-us/library/bb124706.aspx.
Normally the restore of database and log files together with recovery and mount will bring the stores in the Exchange Server instance or th Recovery Storage Group online. However if it is not done automatically then the process of 'hard recovery' can be done by using the following command:
The eseutil.exe utility is located in the Exchange Server \bin folder. Other reasons to run the eseutil.exe utility are problems with the log files restored to the Exchange Server instance or the Recovery Storage Group. In this circumstance, the backup application may report such errors as “Existing log file is not contiguous” or “Existing log file has bad signature.” The following command can be run to force 'hard recovery' and skip the log files, however data in the log files will be lost, so this should only be used as a last resort:
Other issues that may be faced in RSG recovery are failures to properly locate the account or mailbox in Active Directory, typically caused by latency issues between Domain Controllers. It may be necessary to wait several minutes or close and re-open the System Manager for Exchange Server 2007.
Note If Mailbox Log files are missing then a database cannot be recovered over the missing gap. Only a FULL backup together with the complete sequence of following log files can be recovered. In consequence only one backup tool can be used to truncate log files!
The 'Recovery Storage Group' (RSG) feature allows to mount a second copy of a mailbox database - typically a restored one - on an Exchange 2007 Mailbox Server. It was introduced in Exchange 2003. Data can be extracted from one or more mailboxes even from a former backed up version without disturbing the productive databases.
In Exchange 2007 the new 'Database Recovery Management Tool', which is found under the 'Exchange Toolbox Work Center', is used to access RSGs. RSGs may also be managed via the 'Exchange Management Shell' (EMS).
After merge, copy or extracting the required Mailbox data the 'Database Recovery Management Tool' or the 'Microsoft Exchange Troubleshooting Assistant' (ExTRA) may be used to dismount and then remove the Recovery Storage Group. The files in the RSG folder should be deleted after removing the RSG.
Just as with Exchange 2000 and 2003, you can recover an Exchange 2007 Server in a fairly straightforward way. As you may know, we could use the 'DisasterRecovery' switch to recover a dead Exchange 2000 or 2003 Server on new hardware, but with Exchange 2007 this switch no longer exists. Instead, it has been replaced by the new 'RecoverServer' switch, which is similar to the 'DisasterRecovery' switch.
The 'RecoverServer' switch can be used to recover all types of Exchange 2007 Server roles except for the Edge Transport Server role, which uses ADAM and not the Active Directory to store configuration data.
If you’re recovering an Exchange 2007 Server with the Hub Transport Server role installed, and this is the only Exchange 2007 Server with this role installed, it’s recommended you run Setup.com /M:RecoverServer with the /DoNotStartTransport syntax since there’s a few post-recovery steps that should be completed before this role is made active.
When you run Setup with the RecoverServer switch on a new Windows 2003 Server that is configured with the same name as the one that has crashed or is permanently down for some reason, Setup will read the configuration information for the respective Exchange 2007 server from the Active Directory. In addition to applying the roles and settings stored in Active Directory, Setup will (just as when installing an Exchange 2007 Server role without the RecoverServer switch) install the Exchange files and services required for the respective Exchange 2007 server role(s).
The ExSetup.exe programm allows to use the command 'Setup /Mode:RecoverCMS' to recover an Exchange 2007 Clustered Mailbox Server (CCR and SCC).
If a Single Copy Cluster (SCC) is recovered and the Mailbox and Public Folder databases are stored on a Storage Area Network (SAN) then there is no need to restore the databases from backup as long as each node has access to the same Shared Storage Subsystem with the mailbox database files on it.
Description for Using 'RecoverCMS' to recover a Clustered Mailbox Server is located under "How to Use RecoverCMS to Recover a Clustered Mailbox Server" http://technet.microsoft.com/en-us/library/bb124095.aspx
With the new Mailbox Database Portability Feature a Mailbox database can be transferred to any server in the Exchange 2007 organization and with the new Autodiscover Service the Outlook 2007 clients will be redirected to the new server the first time they try to connect after the Mailbox database has been moved and mounted on another Exchange 2007 Mailbox Server.
The name of the new Mailbox database, created with the 'New-MailboxDatabase -StorageGroup <Server_Name>\<StorageGroup_Name> -Name <Database Name>' command, must be identical to the name of the original Mailbox database, otherwise it couldn't be mounted.
See "How to Restore a Mailbox Database to a New Server with Database Portability" under http://technet.microsoft.com/en-us/library/aa998340.aspx.
Our sales team would be happy to assist you!
Browse and discover videos by SEP.