vCenter VM using dvSwitch replication

[stephaneH.]

Hi all,

I'm trying to replicate my vCenter VM using Veeam. In the job config, i have set at destination the direct connection to the target host.
I'm using distributed vSwitch for the vSphere invironnement.

My problem is : when i try to initiate a Failover, the replica VM is started but without network connection.

In the ESXi host event, I can see the network adapter connected then disconnected again and again.

What is the best practice to replicate vCenter VM connected to dvSwitch ? if it possible with Veeam ?

Thanks for your help

[Vitaliy S.]

Hi Stephane,

It could be a stupid question, but have you done network mapping in the replication job settings? Are there any error events on the target ESXi host that are triggered at the same time as adapter connection/disconnect events?

Thanks!

[stephaneH.]

Thanks for the quick answer.

Yes, I have try both way : with and without network mapping to the same dvPortGroup and "duplicate" dvPortGroup.

During the replication, i can see error on the target ESXi. The task "Clone ports" initiated by the vpxuser failed with status "The object or item referred to could not be found."
And just before (and after), the task "Delete ports" is finished without error.

[Vitaliy S.]

Hmm...can you open a support case with our technical team and let them review the log files, as I'm not aware about any known issues regarding dvSwitches.

[michaelryancook]

You may want to try using ephemeral port binding on the dvswitch that the vcenter is connected to. We ran into an issue where we had a complete power outage and the host the vCenter was last on could not boot back up. We re-inventoried it on another host but the host displays "invalid backing" for the NIC on the vCenter VM. The issue is that without vCenter being up the hosts can retain dvswitch port allocations but are not able to allocate a new port on the dvswitch. To do that they require vCenter. Since vCenter is the VM that requires the new port you are stuck. Ephemeral port binding allow the host to allocate a port. In a failover I am guessing this would cause the same situation and ephemeral port binding would resolve. What we did was create a separate dvswitch that used ephemeral port binding and we attached vCenter, the SQL server hosting the vCenter DBs and our domain controllers. Everything else was left on the original dvswitch with static port binding.

I may be off base but hopefully it helps.

[dellock6]

Indeed ephemeral binding can be a quick workaround to avoid invalid backing.

Luca.

[stephaneH.]

I tried with ephemeral, but i've got the error "Processing configuration error: The object has already been deleted or has not been completely created".

Maybe something arround VM hardware version. I try with a VM in vmx-10 version...

I will create a support request.

[Vitaliy S.]

Once you do this, please post your case ID in the topic, so that I could also take a look at the logs. Thanks!

[infraerik]

Hi there - did you ever find a solution to this one?

I have a similar issue trying to setup a replication job with Veeam 11 from ESXi to ESXi of the VCSA appliance so that I can do a planned failover.

[PetrM]

Hello,

The issue was reported approximately seven years ago so I wouldn't rely on the solution in this topic even if there was one. Could you please contact our support team and share a support case ID over here? It would be difficult to find the cause of misconfig over the forum posts.

Thanks!