Object not found. after vm's vmotion

[Nikolay]

Hello,

We are using Veeam B&R 6.5 with latest patches for back up our VMWare 5.0 virtual machines. Backup jobs were running just fine for about a month but at the last weekend we encountered with one issue. One of our hosts in vmware HA cluster was restarted and VM's have been moved to other host, after that these vm's have been moved back to the original host. The veeam job for vms on this host followed by these actions failed with the following error for each vm:

Object [vm name] not found.

After that I have re-added all the vms to the job and Veeam started to backup these vms from scratch - it did not recognized the old full backup for the vms although none of the job settings were changed.

Could you please explain why this has happened and how to avoid such issue in the future?

Thank you,
Nikolay

[veremin]

Hi, Nikolay.

Only adding VMs to the backup job through vCenter Server connection makes backup/replication job vMotion aware. So, may I ask you how these VMs have been added to your job – via vCenter connection or through standalone host?

Thanks.

[Nikolay]

Hi Vladimir ,

Thank you for the reply. Yes, I know that only vCenter added VMs are vmotion aware. But our problem is slightly different.
Our VMs were added to the Veeam job from standalone host.

I'm talking about the case in which VMs were moved to other host with or without vMotion (maybe due to HA restart) and after that these VMs were returned to the original host. Unfortunately, In such case Veeam job fails to backup VMs despite the fact that at the moment of job start these VMs are located on the same host they were at the moment of job creation and previous runs.

Can I avoid this somehow?

Thanks,
Nikolay

[foggy]

Veeam B&R tracks VMs by their unique moref ID (like vCenter does). Every VM re-registeration in the virtual infrastructure (like on moving VM between standalone hosts) changes this identifier causing the VM to be treated as a new VM.

You could use VM containers (folders, resource pools) as a source for the job so it could still pick the VM up and/or map the job to the existing backup file to avoid full backup of this VM.

[Nikolay]

Thank you for the reply,
This thread can now be closed.

[KristiEva]

Hello,

I'm going to be troubleshooting this myself this morning; howver, I imagine that someone has already come accross this behavior so I figured I would post in case there was a known solution that could be shared.

In my environement I have a two Host HA cluster at production and a single Host at my DR site with the production and DR sites with dissimiliar networks. To avoid having to reconfigure my VCServer at the DR site in case of a disaster I point my replication jobs directly at my Hosts.

Yesterday I had to vMotion migrate my VM's to the HA slave while I patched the primary Host. Afterwards I vMotioned the VM's back to the Host that my Veeam jobs point to; however, after they were moved back Veeam no longer "sees" them and my jobs fail with:

Task failed Error: Object "SERVERNAME" not found.

In the replication job > Virtual Machines tab for each failing job the VM shows a size of 0.0KB (I assume because it doesn't actually see the VM).

Is there a way to reset the Veeam job so that it can still see the original VM on the Host the job was configured to look at after the VM is migrated back?

[KristiEva]

I think I found my answer:

http://forums.veeam.com/vmware-vsphere- ... 15492.html

Is the information in the above thread still accurate for vSphere 5.5 and Veeam v7?

[foggy]

Yes, since VMs tracking mechanism hasn't changed, the behavior you're observing is totally expected. Thanks.