Hot Add, attaches Disk being backed up to incorrect Veeam VM

[a1lenski]

Hi All,
I have a very odd issue with a backup running from Veeam 6.5.0.109,

First Time the Job Ran (Full) The source VM was backed up using Hot Add, successfully.
Second Run (Reverse Incremental), At the Hott Add stage, the disk form the source VM is added to the Replica VM of the Veeam VM running the Job, as opposed to the Veeam-VM itself, this causes the backup to fall back to Network mode and Run at a slower transfer rate.

I am running 5.0 VMware, Both The Veeam-VM and its replica Veeam-VM_Replica are within the same Virtual centre.

Any advise would be greatly Appreciated.

[dellock6]

Usually Veeam-VM_Replica is powered down and kept only to failoveer in case Veeam-VM crashes, so how can it happen the job do hotadd to this VM???
Maybe you created the replica VM using default settings, and you ran replica while the VM itself was doing hotadd-based backup? I've seen this behaviour while replicating some proxies: they were doing hotadd backups, so they had their own disk (scsi0:0) and several other disks coming from the hotadd. Replica tried to replicate also all these disks. Configuring replica to replicate only scsi0:0 solved the issue... Do not know however if this is your case...

Luca.

[Vitaliy S.]

Usually Veeam-VM_Replica is powered down and kept only to failover in case Veeam-VM crashes, so how can it happen the job do hotadd to this VM???

I should admit that this is a super valid question Looking forward to some further clarifications on this too. Btw, do you see the same behavior for other VMs being backed up?

[a1lenski]

HI Guys,
Thanks for the responces, It does seem very strange. I am pretty sure the Replica-VM was produced when no backups were running. It is in a powered off state incase of a crash, since I am not 100% on its stae when created I will blow away the replica VM and re-create.

I have only seen this with one particular Job so far but I was not watching while the other scheduled Jobs run as I am investigating a slow backup issue for that particular Job, I created the Job after updating Veeam to current version. I changed how the repository was being accessed from a CIFS connection to an iSCSI one provisioned directly to the Veeam-VM using the MS initiator. This has improved the backup speed. I can check and advise if this occurs with any other jobs.

I am thinking that somehow the GUID for the VM is wrong / confused and when Veeam is calling the VMware backup API it is selecting the wrong machine, the power state of the VM shouldn't matter when adding or removing disks to it, as Hot add is just attaching the VMDK to the VM using the Vmware API.

Thanks in Advance

[foggy]

As far as I know, it is not possible to have two VMs with the same UUID in vSphere environment, so this should not be the case. However, VM UUID conflicts can be easily identified with the help of Veeam ONE, for example (it will report the "VM instance UUID conflict" alarm in this case).

I would recommend to investigate this with our technical support, probably debug logs will reveal the reason of this behavior.

[pcrebe]

Hi,
I've got the same problem.
Veeam sw hot adds disks to proxy replica servers, that are turned off, using failing over to network mode. I need deselect the Automatic proxy servers option and specify the proxy servers i want use.

I've EnableSameHostHotaddMode active for the nfs locking

[veremin]

What has been said above still applies here - the situation when VB&R tries to attach disks to powered off VMs looks unexpected. So, feel free to open a ticket with our support team and let them investigate it directly. Thanks.

[pcrebe]

Veeam Support - Case # 01014285