Backup Clustered SQL Servers via fibre channel

[mniechcial]

Hello,

How should i install Backup Exec 12.5 to backup databases located on cluster sql via fc. I have Primary server and I want to install on both nodes of cluster BEWS (as secondary media servers) to use comunication with robotic library via fc. I'm not sure if this will work. How should I configure backup jobs on both nodes, to get backup even if one of nodes will crush? One of solution is script on second node which checks if first node is available - if it is available, second node doesn't start the backup job. Maybe exists an easier way to do this.

[JoaoMatos]

Hi,

You can always install BE in Cluster mode, though I do not fancy that option.

You can install BE Media server on one node, and run a backup job to the virtual server. 

Running the backup on the virtual server guarantee always back up your data even if one node fails

The only problem with this solution is if the node where the BE is installed fails.

[Hywel_Mallett · Accepted solution]

It sounds like the best solution for you is to use the Central Admin Server Option. This coordinates jobs between media servers.

You'll need two additional Backup Exec licenses, for your cluster nodes. As you want to share your tape library, you'll also need three Shared Storage Option licenses.

[CraigV]

...just to add in to what Hywel suggested. If you use the SAN SSO option (SAN Shared Storage Option) on your primary server (it is licensed!), this will allow you to share your FC-attached library. Your full version of BE will then get installed on your cluster nodes, and you run the SAN SSO wizard on them to share the library.

If you don't have an FC-attached library, this won't work...your backups will always go over your LAN!

[CraigV]

Hi,

Any news here?

Thanks!

[mniechcial]

Sorry for this huge delay.

I installed full BEWS with SQL agent on both cluster nodes (eg. SQL1 and SQL2). Both servers are backing up via  fc (System partition, System State and Shadow Copy). On both nodes I created a job which backs up databases located on cluster (CSQL) but on SQL2 this job is hold - will be use only in case of crash of SQL1. Job rate shows that this backup job is done also via fc. What do You think about this configuration?

[CraigV]

...that will work well, albeit manually.

You are able to run a cluster configuration wizard if I remember correctly. So as your cluster moves over to the inactive node, it moves BE as well.

If you download the BE 12.5 Admin Guide and read Chpt. 18, it will give you better insight. Give it a read, and post back the results...