Retention set for TDP Vmware data

[Silapen]

On 8.1.x version retention sets were introduced as new feature for long term data keeping. Has anyone used retention sets for node that holds Vmware backups (created using TDP for Vmware)?
Is it supported?
How to restore specific VM?

Thanx in advance.

 

[Silapen]

See a lot of views on my original question, but no answers, so let me expand the question a bit. What is your strategy for long time retention of TDP for Vmware data? Do you create a separate node with different copygroup/mgmt settings and then run a backup to that node manually?
Do you use exports of the node data to tape?
Some third option?

Best regards.

 

[Mita201]

Can't think about good reason for storing copies of VMware data (that has been regularly backuped) for some long time.
I may think about doing it under separate nodename, but just not sure how it will affect CBT record of previous backups. Maybe it will break the chain, so you will do another full backup next time whemn regulat backup occurs.

 

[marclant]

It is supported and you would restore it as any other backup. Retention sets don't change the backup data, it only affects how it's retained. So you could create retention sets of month end or quarter end backups of certain VMs if you have a business requirement to do so.

 

[P-to-H]

For long-term backups, we extract the VM as a file and save it by archiving.

This means that we are independent of TSM4VE and are guaranteed to always be able to restore the VM. It is then up to vmWare whether such an old VM is still supported.

 

[DanGiles]

I am curious to know the type of data on these servers, and the reason for long-term retention. If they contain a database or other well-known/supported application, then the retained backups should be done through there. If its a custom application, then I can see archiving the whole machine whenever there is an upgrade to that application, or some structural change. If its specific data, then I would just archive that data. Either way, I would give it a separate node name so that I can place the different data sets into different storage pools. (Think DR!)