hot add failures?

[danswartz]

I am running Veeam B&R 6.5 on my windows 7 workstation. This means all of the replication traffic for the 5 VMs goes to my workstation, then back to the vsphere host, which then has to send it back out to the NAS filer. I have a windows server 2008 VM, so I set it up as a backup proxy. All of the guests except w2008 do in fact try to do hotadd, and they all have messages in the log that they are using hotadd for source *and* target proxies. About 50% of the time though, later in the log after the 'creating helper snapshot' message, it says hotadd is not supported for this disk, failing over to network mode... Immediately after that is a CBT SOAP fault message (which is not helpful - for me anyway). Since CBT is disabled for that run, the replication takes a LOOOONG time I've been running this since mid-morning (hourly replication), and it seems kinda random. e.g. it isn't every VM that fails, nor is it the same VM each time. Once it was 2 of 5, the next time 1 of 5, the next time 2 of 5, but the problem moves around. Any ideas? Thanks! One other thing I noticed on the 2008 console - several messages about 'you need to format this disk before you can use it'. That seems sporadic, and I have no idea if it is related or not. Additional data point: looking in the win2008k logs, I see messages from the plug&play manager about 'vmware scsi disk disappearing without being prepared for removal'.

[danswartz]

I apologize if the above post is deemed to violate the rules of posting. My replications are in fact working, just not as efficiently as possible. If you wish, I will open a support case...

[Gostev]

Hi Dan, yes please do... your issue is not a common/known issue, and we cannot investigate any issues without the debug logs. Thanks!

[danswartz]

Will do, thanks.

[danswartz]

Uploaded logs. Before anyone at veeam spends time on this, want to make sure I need this. As I said in the OP, the point is to avoid unnecessary network traffic to&from my workstation. If I continue to use the win 2008 server, but force the proxy to use network mode, most of the traffic will stay localized to the vmware server itself - only the final leg (the disk writes to the NAS filer) will leave the box. Does this make sense? (this is case #00473051).

[Gostev]

Yes, this makes perfect sense.

[veremin]

Also, I’m wondering whether it can be the problem with particular proxy server. Are using only one proxy server in your environment? If so, it might be worth temporarily deploying additional one, switching backup workload on it, and seeing whether the problem reproduces or not.

Thanks.

[danswartz]

That is the only proxy, yes. It's almost got no load on it at all, and the host itself is relatively idle, so I doubt it's load-related, but who knows? I did switch that proxy to network mode, and am getting good performance with no glitches, so I think I'll call it a day, if you don't mind. It's most likely some kind of vsphere glitch (running 5.1), but I don't really have the time or energy to chase unnecessary problems. Thanks for the suggestion, though!