How I Successful Failed Using SCR

One of my recent Fun and exciting tasks at work was to fail from an Exchange 2007 CCR cluster to an SCR failover machine. There are two ways to configure the SCR failover:

  • Standalone Server    Configure the failover SCR server as a standalone mailbox server. When you will fail to it you run Move-Mailbox –ConfigurationOnly for all of the users in the failover database to tell users that they have moved to a new server. When Outlook 2007 connects to Exchange AutoConfigure will reconfigure Outlook to point to the new SCR server.
  • Passive Cluster Node   Configure the SCR failover sever as a passive cluster node. When you fail to it you run Setup.com /RecoverCMS and seize the Cluster name. There is no further user configuration needed after this.

For my test I used the second option and configured the SCR failover server as a passive cluster node. This was simple to do, all I did was pick Passive cluster node during setup and then configured SCR and I was good to go. The following sections cover the steps that I took to Fail to SCR, Fail Back to CCR then reestablish SCR. At the end are some of the issues that I ran into that you might want to think about as well. I wrote the follow sections as notes to help me know what to do when I did this, and I refined them as I failed over and back. The server names are the real server names that I used, the CMSIPaddress was changed to protect the innocent.

This is a long work of nerdness lacking in pictures. For those interested in the steps and procdures click on Read the Rest of the Story link below.

Move from a CCR cluster to SCR

1. Verify that the SCR replica is healthy
2. Shut down the cluster nodes ( take them offline or unplug them as you see fit)
3. Restore the SCR replicas by running the following Exchange management Shell Commands:

Restore-StorageGroupCopy –Identity “MAPI1\First Storage Group” –StandbyMachine EXCHANGE-SCR1 -Force
Restore-StorageGroupCopy –Identity “MAPI1\Second Storage Group” –StandbyMachine EXCHANGE-SCR1 –Force

4. Recover the CMS node to the SCR server, using the Cluster IP of the CCR node by running the following command:

setup.com /RecoverCMS /CMSName:MAPI1 /CMSIPAddress:10.10.10.10

5. Mount the databases by running the following Exchange Management Shell commands:

Mount-Database –Identity “MAPI1\First Storage Group\Mailbox Database”
Mount-Database –Identity “MAPI1\Second Storage Group\Public Folder Database”

6. Bring down the CAS + HUB servers and add their IP addresses to the DR CAS server

Move from a Passive SCR Cluster node back to the CCR Cluster

1. Bring both EXCHANGE-MB1 and EXCHANGE-MB2 back online
2. Remove the Cluster resources from EXCHANGE-MB1 by running the following command:

setup.com /ClearLocalCMS /CMSName:MAPI1

3. Create an SCR replica of the database on the EXCHANGE-MB1 sever by running the following Exchange Management Shell commands:

Enable-StorageGroupCopy “MAPI1\First Storage Group” –StandbyMachine EXCHANGE-MB1 –ReplayLagTime 0.0:0:0
Enable-StorageGroupCopy “MAPI1\Second Storage Group” –StandbyMachine EXCHANGE-MB1 –ReplayLagTime 0.0:0:0

4. Reseed the database on EXCHANGE-MB1 by running the following Exchange Management Shell command from the destination server:

Update-StorageGroupCopy –Identity “mapi1\First Storage Group” –StandbyMachine Exchange-MB1

5. Dismount the databases on the SCR server by running the following Exchange Management Shell commands:

Dismount-Database “MAPI1\First Storage Group\Mailbox Database”
Dismount-Database “MAPI1\Second Storage Group\Public Folder Database”

6. Restore the databases on the EXCHANGE-MB1 server by running the following Exchange Management Shell commands:

Restore-StorageGroupCopy “MAPI\First Storage Group” –StandbyMachine EXCHANGE-MB1
Restore-StorageGroupCopy “MAPI\Second Storage Group” –StandbyMachine EXCHANGE-MB1

7. Recover the CMS node on EXCHANGE-MB1 by running the following command:

setup.com /RecoverCMS /CMSName:MAPI1 /CMSIPAddress:10.10.10.10

8. Rebuild EXCHANGE-MB2 by running the following Exchange Management Shell command on the Exchange-MB2 server:

Update-StorageGroupCopy –Identity “Mapi1\First Storage Group”

Reestablish the SCR replica on the Passive SCR node

1. Remove the Cluster resources from EXCHANGE-SCR by running the following command:

setup.com /ClearLocalCMS /CMSName:MAPI1

2. Recreate the SCR replica by running the following Exchange Management Shell commands:

Enable-StorageGroupCopy “MAPI1\First Storage Group” –StandbyMachine EXCHANGE-SCR –ReplayLagTime 0.0:0:0
Enable-StorageGroupCopy “MAPI1\Second Storage Group” –StandbyMachine EXCHANGE-SCR –ReplayLagTime 0.0:0:0

3. Reseed the database on EXCHANGE-SCR by running the following Exchange Management Shell command from the destination server:

Update-StorageGroupCopy –Identity “mapi1\First Storage Group” –StandbyMachine Exchange-SCR

Lessons learned / Mitigated Issues

Things that had to be dealt with / fixed to make everything work

  • The –force option on the Recover-StorageGroupCopy command put the databases in a dirty shut down state that required me to run eseutil /r /a. I tried to run without the –force when possible.
  • I added the IP Addresses of the live CAS + HUB servers to the failover server to avoid DNS changes – we did this because during the test we turned off the active CAS+HUB servers.
  • The receive connectors were disabled on the failover CAS+HUB server I had to enable them before inbound email would work – You need to think about Sending and receiving from your Failover severs and how that will impact things
  • The failover CAS+HUB server was not listed as a source server for the outbound SMTP connector I had to add that before outbound email would work – You need to think about Sending and receiving from your Failover severs and how that will impact things

Things that need to be fixed later

  • Valid SSL certificate needs to be installed on the failover CAS+HUB server
  • OWA folder redirection needs to be enabled on the failover CAS+HUB server if it is enabled on the primary server
  • A script needs to be written for the SCR failover so we don’t have to do it by hand next time
Related Posts with Thumbnails

About Kevinm