NOTE: When we say "Exchange 2003 Standby Cluster" we are referring to the configuration discussed here:
http://technet.microsoft.com/en-us/library/d2492381-e9d5-4827-b7ff-7ac614ccc36d.aspx
With Exchange 2007 we have the ability to recover an Exchange cluster with the exsetup /recoverCMS switch. This is similar to the functionality we have with Exchange 2003 with the setup /disasterRecovery switch with a standalone system.
Note: In Exchange 2007 an Exchange Virtual Server is now called a Clustered Mailbox Server (CMS).
The steps to recover an Exchange 2007 Single Copy Cluster (SCC) or Cluster Continuous Replication (CCR) are the following:
For the Single Copy Cluster (could also be explained as "shared storage cluster"):
- Build an MSCS Cluster. When configuring the cluster it is required to have the same structure as the original shared disks in the cluster (same logical disk names). /RecoverCMS will pull this information from Active Directory, and if the disk structure does not match the original server, recovery will be blocked.
- Install the passive Mailbox server role on the node on which you intend to recover your clustered mailbox server.
- To recover the clustered mailbox server, change to the bin directory under the Exchange program files. By default, the installation file location is <systemdrive>:\Program Files\Microsoft\Exchange Server\bin. Run the following command: exsetup /recoverCMS /CMSName:<CMSname> /CMSIPaddress:<CMSIPAddress>
- If you have multiple CMS’ in the SCC, run the previous step above on a separate node to install another CMS, repeat until all active nodes are installed.
- Restore databases if they are not already available on the storage system. If set, remove the Do not mount this database at startup setting from the Exchange databases by running the following Exchange Management Shell command: Set-MailboxDatabase <restored_database> -MountAtStartUp $true
- Install the passive Mailbox server role on other passive nodes.
For the CCR Cluster:
- Build an MNS Cluster. When configuring the cluster it is required to have the same structure as the original local disks in each node of the cluster (same logical disk names). Recovery will be blocked if the disk structure does not match the original server.
- Install the passive Mailbox server role on the node on which you intend to recover your clustered mailbox server.
- To recover the clustered mailbox server, change to the bin directory under the Exchange program files. By default, the installation file location is <systemdrive>:\Program Files\Microsoft\Exchange Server\bin. Run the following command: exsetup /recoverCMS /CMSName:<CMSname> /CMSIPaddress:<CMSIPAddress>
- Restore databases if they are not already available on the storage system. If set, remove the Do not mount this database at startup setting from the Exchange databases by running the following Exchange Management Shell command: Set-MailboxDatabase <restored_database> -MountAtStartUp $true
- Install the passive Mailbox server role on the other node in the cluster.
- After restoring a CCR cluster, replication is suspended. You must reseed the replica and resume replication.
For both cluster types you can create the new cluster with all new IPs and a new cluster name, but you must use the original CMS name. This cluster can be in a separate location (subnet) from the original cluster. This means you can have a cluster with the passive mailbox role installed and once you are in need to recover you can run the exsetup /recoverCMS switch to recover the nodes and then restore the databases. It is not possible to take an non-clustered mailbox server and recover as a SCC or CCR CMS. It is also not possible to have a SCC recovered as a CCR cluster or vice versa. It is however possible to use the database portability feature to move the databases between different types of clusters, if needed.
In addition, splitting clustered mailbox servers from a single cluster into many clusters or vice versa is also not supported as part of clustered mailbox server recovery.
- Matt Richoux
