CBT and host crash

[mkretzer]

Hello,

For the first time in years we had a host crash (driver issue with the internal raid adapter which we only use for booting - we had to reset the system because we could not manage it anymore). HA did its job good and fast.
Still I wonder: what about CBT? Is it protected in such situations or can it be silently corrupted (aside from the bugs in older versions)?

Markus

[Vitaliy S.]

Hi Markus,

CBT gets reset only when you re-register VM in the inventory. How did you run the failover when fixing the issue with the original host? Do you run SureBackup jobs to verify consistency of your backups?

Thanks!

[mkretzer]

Hello,

we did not have to re-register, HA rebooted on another host. That means we have to verify every database, every file on every VM which ever was on a failed host?

Markus

[vClintWyckoff]

No, you don't need SureBackup, I think Vitaliy misunderstood. If you suspect CBT issues, there's a few options.

1.) Veeam will still perform an incremental backup by using our own filtering mechanism to determine the incremental changes.
2.) You can reset CBT on the affected VMs through PowerShell as per this KB

[mkretzer]

The question is rather: should i suspect CBT after a host failover?

[Gostev]

No. VMware automatically invalidates CBT data in case it is inconsistent, which by the way also means your next incremental backup will take much longer, as without CBT data Veeam would have to read the entire VMDK to determine all changed blocks. This failover happens automatically and transparently though.

[Zew]

Good answer Gostev, glad to see you are still active on the forums

[mkretzer]

So we did a dbcc checkdb of one of our biggest databases which was on the crashing host and the database only shows the same 3 inconsistencies it shows for 8 years or so . Nice!