CBT data is invalid on Veeam: should i let it fail over?

[hasayeret]

So here's a question to the Veeam wizards around :

My CBT got corrupted on my onsite backup about 2 months ago. When I replicated the data offsite I received a message saying that the CBT data is invalid and it's failover over to the legacy incremental backup. 51 hours was the time it took to complete.

Last week I had another corruption in my CBT but before I started the offsite replication I removed the replica files from Veeam so it actually created fresh new files. 61 hours was the time it to complete this one: 10 hours more than what it took 2 months ago.

The processing rate was the same, the bottleneck was the same and the size of the VM was the same and both jobs used SAN for source proxy and HOTADD as a target proxy. HOWEVER, in the 51 hours job the read on the hard disks was A LOT faster.

Did I need to keep the files without deleting them in that 61 hours job or do you think that there's another reason?

very very curious about this one...

Thanks!!!

[PTide]

Hi,

HOWEVER, in the 51 hours job the read on the hard disks was A LOT faster.

How big was the difference? Were the VMs under the same load in both cases? May I ask you what's your ESXi version?

Thank you.

[foggy]

I'd also recommend to investigate the reasons of CBT corruption that caused the discussed behavior in the first place.

[hasayeret]

I'm running ESXi 6.0.0 build 3247720. The CBT got corrupted because the server was rebooted by a mistake in the middle of a job which is not going to happen again. The VMs were under the same load and the difference was pretty minor or at least minor enough for me to not notice

[veremin]

I think it's worth opening a ticket and find out the root cause via logs' investigation, as situation looks strange given the same conditions. Thanks.