Forum Discussion

MitchR's avatar
MitchR
Level 6
13 years ago

Agent Trust & Cluster issues

I'm using Backup Exec 2010 R3, SP2.  One of the computer I'm backing up is a Windows 2008 file server cluster, consisting of 2 nodes.

I installed the agent on bath cluster nodes.  I setup the agent's trust on both nodes.  The cluter resource I'm backing up has it's own IP, and points to either of the nodes, depending on which hardware node owns the resource.  Note that I'm using Windows File Server resources, not a VM or any hosted OS. Standard Windows file clustering stuff.

Well, the backups only work when the file server resource is on the first node.  When we fail over the file server resource to the 2nd node, BackupExec start complaining that the trust relationship has not been setup.  We fail back to the first node, and backups work again.

It seems that since this whole agent trust security was implemented, BE can't handle file server clusters.

Anyone have a fix or workaround?

 

2010
Agents
Backing Up
Backup and Recovery
Backup Exec
Error messages
Other-New Agent or Option Request
Remote Agent for Windows Servers

4 Replies

Sort By
  • pkh's avatar
    pkh
    Moderator
    13 years ago

    Have you tried pushing out the RAWS to the 2nd node again to re-establish the trust relationship?

  • MitchR's avatar
    MitchR
    Level 6
    13 years ago

    Yes - and that makes the 2nd mode work.  However it then breaks the trust on the first node.

    It's like BE thinks the IP address of the clustered file server is a single server, and can't handle having two different trust certificates, depending on the node it's hosted on.

     

  • pkh's avatar
    pkh
    Moderator
    13 years ago

     

    You should upgrade to SP2 and all the subsequent hotfixes.  Make sure that you have Hotfix 180429 which fixes trust problems.

    Remember to push out your agents after the updates.

  • CraigV's avatar
    CraigV
    Moderator
    13 years ago

    Mitch: Shouldn't be the case. There is a cluster configuration wizard that you'd run on both those nodes, and you'd obviously only select the virtual node.

    Maybe also look at running the wizard again...