Changing Backup Policy from Incremental to Periodic Full

dcharles226

ADSM.ORG Member
Joined
Apr 2, 2015
Messages
37
Reaction score
1
Points
0
Hello all,

We were wondering if it was possible to change our backup strategy for a small set of VMs that are currently being backed up daily. Currently, they are being backed up as Incremental Forever, but we want to change that to Periodic Full. This means we want to basically take one Full Weekly Backup, and then six daily Incremental Backups each week.

Is there a way to migrate the filespace for each VM so that TSM understands that the backup strategy has changed? I've already tested this on one VM by deleting the filespace from the Data Mover node and starting a new fresh Periodic Full backup, but I am getting errors after the backup has run:

TSM return code : 115
TSM file : vmbackvcm.cpp (218)
11/01/2016 00:38:18 ANS5250E An unexpected error was encountered.
TSM function name : vcmLogger::trace
TSM function : CacheManager::remove_element: Failed to save element: Path: '/tmp/tsmvmbackup/fullvm/CDF_Local/VM/Hard Disk 1' Job: 1000001 MB Number: 1

We are running TSM Server 7.1.4 on Red Hat 6.2 and TSM for VE 7.1.4 on CentOS 6.5.

Thank You for your help
 
urrently, they are being backed up as Incremental Forever, but we want to change that to Periodic Full.
Curious to see what advantage you see in doing that? The number of recovery points would be the same one way or the other. You just use more storage with the weekly full + daily incremental. The only advantage I can see is if you backup to tape, you'd get better restore performance because the the data would be together with the weekly full.

Is there a way to migrate the filespace for each VM so that TSM understands that the backup strategy has changed? I've already tested this on one VM by deleting the filespace from the Data Mover node and starting a new fresh Periodic Full backup,
There's no way to change the filespace, so you have to do like you did, and either delete or rename the filespace, so a new one is created with the current backup strategy.
TSM return code : 115
TSM file : vmbackvcm.cpp (218)
11/01/2016 00:38:18 ANS5250E An unexpected error was encountered.
TSM function name : vcmLogger::trace
TSM function : CacheManager::remove_element: Failed to save element: Path: '/tmp/tsmvmbackup/fullvm/CDF_Local/VM/Hard Disk 1' Job: 1000001 MB Number: 1
That's unrelated to your change, because all your change did was basically start fresh. I'm not sure what's the cause of the error.
 
Thanks for the quick response.

So that is correct, the point of changing the backup strategy is that we want to be able to restore faster from tape. We recently did some restore tests on some random VMs and the restore speed from tape was horrendous. It was then suggested that we change to Periodic Full in order to reduce restore time from tapes.

I'm still not sure why we got that error, since I delete the filespace related to that VM in order to start a fresh backup set. In the worst case scenario, I might have to create an entirely new data center node and backup the Period Full VMs to that one.
 
So after completely deleting the filespaces for the VMs we wanted to change to Periodic Full, it looks like the backups are working now as intended.

The error message I mentioned in my original post seems to be related to me forgetting to remove that VM from another schedule that it was associated with. Once I removed the VM from that other schedule, it backed up just fine.
 
Back
Top