File Server Backup in a Cluster Differentials Run a Full When Cluster Fails Over

[Les_Denniss]

Hello hope someone can advise!

We have a two node file server cluster ( Windows 2008 R2) backing up a number of drives using Differentials and Synthetic Full backups. When a drive is moved over, or fails across to the other node, the differential runs a full backup, I understand this is due to the archive bit being reset!

In NetBackup, under Windows Client > Client Settings > Incrementals - we have it set 'Based on archive bit' should we set this to use 'Timestamp'? or could this cause other issues?

Or can you recommend a better solution!

Thanks in advance!

[SymTerry]

That usually doesn't need to be configured.  Also Just to make sure, you have a client installed on each of the nodes and a individual policy for each Node of the cluster?

look at the Configuring NetBackup Policies for Clustered clients doc as well.

[RiaanBadenhorst]

Post the output of bppllist "policy name" -U

And tell us what the node name and virtual names are.

[Les_Denniss]

Thanks SymTerry , Riaan for your response.

To be clearer on our architecture we have 7.6 client installed on both nodes. we have 8 drives on one node and the other passive. Each drive letter has its own backup IP so in theory we can restore direct regardless what node it is on and should follow if it fails over to the other side. We then have a seperate policy for each drive so the backups can be staggered. This issue occured when someone moved one of the dirves to the other node during the week and that drive run a full.

[Les_Denniss]

Hello I hope someone can help. It only seems to be one drive that keeps running a full at the moment - does anyone have any ideas? would setting to timestamp make any difference? - thanks!

[AndreaBolongaro · Accepted solution]

Check if you have configured "Collect true image restore information" in the policy.

If you have activated it, that's normal, because TIR info are stored on local disk by default.

You may also have a look in job datails on activity monitor to investigate.

[Yasuhisa_Ishika · Accepted solution]

I agree with Andrea. Synthetic backup requires TIR, and TIR info is kept on each client. Once service is failed over, TIR seems like cleared, and NetBackup takes all the files like full backup.

[Les_Denniss]

Thanks Andrea / Yasuhisa for you input on TIR

I ran a manual synthetic full after the drive failed to the other side, I thought this would be the new reference point and differentials would continue as normal on the other node!

I have run a new standard full on the new node to create the new reference point and see how the differentials run - now that I am aware I will look out for this in the future!

Thanks everyone for your input - much appreciated.