Relication Server/Proxy - Replica VM Windows Failed to Start

[markfellows]

I’m just conducting some replication tests using Veeam v6.1 located on VMs on 2 Esxi servers using local storage. Veeam B&R is installed on a VM on each of the Esxi servers which act as both a Veeam server (local SQL Express database) and as a proxy for the other Veeam server to provide Hot Add functionality.

There is a job on each of the servers which replicates a VM from one host to the host on which the Veeam server resides . The initial replication jobs on both servers work fine, as do the subsequent replication jobs.

An issue occurs when these jobs are run concurrently. After the replication is complete (which completes without errors) one VM replica boots fine, the other produces the error below:

“Windows failed to start. A recent hardware of software change might be the cause. To fix the problem:”

Full error details : http://www.pctechbytes.com/windows/wind ... art-error/

Obviously this issue should be repairable from a Windows point of view but I just curious as to what may be causing this in the replication process.
The source VM is fine and re-boots fine as does the Instant Recovery version of the VM.

Has anybody seen similar issues before or are aware of any technical restrictions that would prevent this setup from working (server/proxies processing concurrent jobs).

Could it be down to the specification of the proxies/servers? Both utilise the same NAS path on the repository for the Replica Metadata, could this be an issue?

I’ve got no issues with running these jobs independently I’m more inquisitive than anything else.

[Vitaliy S.]

Hi Mark,

I believe this issue is caused by the fact that you're running a replication job that processes the VM with disks that were hot added by another replication job.

Be aware that by default target backup proxy tries to use hot add mode to populate VM replicas on the destination host. Obviously, this might cause some issues in your scenario, as hot added disks might not be present in VM replica configuration file.

To resolve this issue please force your target proxy servers to use network mode, this procedure should be done for each backup server.

Thanks!

[markfellows]

Thanks Vitaly,

Will do, think that makes sense. The strange thing is though the other VM replica works fine (which backs up in exactly the same way but from the opposite server). It also only happens when the replication jobs on the two Veeam servers run concurrently.

[foggy]

The job that manages to be the first to make a snapshot should complete fine, while the second starts to process the VM with disks already hot added by the first one and could have issues.

[markfellows]

Alexander,

They're not processing the same VM or host, below is how I've created the jobs:

Job 1 - Veeam Server (A) -> Replicates VM (B) on Host (B) to Host (A)
Job 2 - Veeam Server (B) -> Replicates VM (A) on Host (A) to Host (B)

I start Job 1 a few seconds before job 2 but it's Job 1 which seems to have the disk issues when I attempt to start the VM Replica on host A.

Alexander & Vitaliy could the hot add issues you mentioned in the previous posts be the issue in this scenario?

[foggy]

Yes, I'm talking about exactly the same scenario.

[markfellows]

Ok guys, thanks for clearing that up.