Backing up a Microsoft Failover cluster resource
I have a couple of questions about running Backup Exec 2010 with a Microsoft Windows 2008 Failover Cluster.
If I can explain our setup first. We have four servers that are joined to a domain and they’re all running Windows 2008 R2. There is a Dell iSCSI SAN that provides the shared storage. Let’s call the Windows servers F1, C1, C2 and B1.
- F1 is a file server. Nothing special.
- C1 and C2 are nodes in the Failover cluster. They are running a couple of applications on the cluster:
- IIS which uses a cluster disk called U
- SQL Server which has two cluster disks; F and X
- B1 has Backup Exec 2010 installed on it.
Each machine has a network connection for corporate traffic and a network connection for the backup network. The CORP network subnet is 220.127.116.11/24. The BACKUP network is 18.104.22.168/24.
The IIS service has an IP address on the CORP network, say 22.214.171.124, and an IP address on the BACKUP network, say 126.96.36.199. The SQL server is setup in a similar fashion – 188.8.131.52 for CORP and 184.108.40.206 for BACKUP.
We’ve pushed the Windows Remote Agent out to all machines from B1. We are able to perform (from B1) backups of the F1 server without problem over the backup network (setting the network interface in the job properties screen). Our problem comes when attempting to backup the cluster resources on the clustered services. We are able to see the clustered disks U, F and X under the cluster names, and we’re able to select them for backup. The problem is that when we run the job using only the backup network we get a message that says the machines may not be on the same subnet. They are. If we allow the backup to run over any available interface then the job does run without problem.
Given this, and my lack of experience with Backup Exec, I thought that it may be that the agent was not bound to correct IP address, but the fact that F1 seems to backup ok seems to go against this.
So, like I said, I have a couple of questions.
- Does anyone know why we are not able to backup the cluster resources over the backup network, but we are over the corporate network?
- Is this even the way that we should be backing up the disks?
If the answer to question 2 is that we should be doing something else I suspect that I won’t be able to change it. I’m told that the design has been fixed and that the cluster resources must be backed up via the cluster, however it’d be useful to know how it should be done.
Thanks for taking the time to read this, and for any answers that you can give.