Full Backups and Changed Block Tracking

[KiwiJJ]

Hi,
We run our backups as a full backup every day. At the moment Changed Block Tracking is enabled on all of the backup jobs. This creates a little bit of overhead on the VMware host as it writes to the *ctk.vmdk file as well as going through a stun/unstun cycle as it allows a change tracking filter to be inserted into the storage stack for that VM when certain operations such as Veeam backups start/finish. I am noticing that this "stun" cycle is causing our VM's to pause for up to 9 seconds when the Veeam backup jobs start / finish.
As I am doing full backups is Changed Block Tracking of any benefit ? if not and I turn the option off in the Veeam backup job will Veeam turn off CBT on the VM or will I have to do that via the configuration parameters of the VM ?

I have notcied in the summary of the backup jobs where CBT is enabled that a lot of VM's are reporting that it is not being used and some are which I find a bit odd.

thanks,

JJ

[Sethbartlett]

You will want to turn it off in Veeam and in VMware. If you turn it off in Veeam, your problem will still persists, and if you only turn it off in VMWare, Veeam will turn it back on. You're not going to see any benefit to CBT as it is not used during a full backup.

[KiwiJJ]

Hi,
After turning off CBT in both VMware and Veeam the server no longer hangs / pauses when doing the commits back to disk at the end of the Veeam job which means I have happy users.
So definately seems as if CBT was causing the problem.

regards,

John

[Gostev]

Are you using NFS storage?

[mstoeckle]

Hello Gostev,

we experience similar problems here: Unstun times of 30 - 180s on an NFS NetApp.

Did your question about using NFS imply any remedy for those problems especially in an NFS scenario?

Thanks + Greetings, Martin

[Vitaliy S.]

Hello Martin,

Please take a look at the existing topic to workaround your issue: Snapshot removal issues of a large VM

Thank you!