[Feature Request] Improve Direct ESXi Host Communication

[BrianS_WincVA]

Reference Support Case 06169996

Veeam hangs for about 90 minutes if vCenter is down and you try to do a replication or failover even if the job specifically references the hosts directly.

Steps to reproduce:
1. Add a vCenter server to the infrastructure
2. Add at least 2 ESXi standalone hosts by their IP addresses (you get a conflict error if you try to add using DNS names)
3. Configure a replication job pointing the source and destination to the standalone hosts. In my scenario, I am trying to protect the vCenter VM itself, however you will also see from the details of the support case I tried another VM as well and the same issue will occur.
4. Start the replication job and it will complete successfully. You can observe vCenter as well as the ESX hosts and see all the tasks run on the standalone hosts and not vCenter - exactly as configured and expected.
5. Shutdown the vCenter VM
6. Start the replication job and it will hang at "building a list of VMs" for about 90 minutes, then connect to the ESX hosts and the job will proceed normally.

In the support case, you will find a Broker service attempting connection to vCenter nonstop for 90 minutes before moving on to the ESX hosts. Since vCenter is so critical to Veeam's operation, and it is known that vCenter HA is not supported, we need an easy way to failover and failback vCenter during a DR event.

Thank you for considering this request. Let me know if you have any questions.

[Andreas Neufert]

Hi Brian, thanks for sharing. I used the same scenario in my demo lab for years where the vcenter was practically always down without any issues, so I think this is a bug or is caused by some specific configuration.