CDP out-of-memory on proxy leads to a corrupted vm

[mcz]

Dear PMs,

I'm creating this thread a bit offset to the event. It all started when I upgraded to v12, since then 1-2 times per week I had the problem that veeam wasn't able to create long-term-retention points for some vms (CDP). There was no logic and it has always hit different vms and obviously there was also no recovery from that - the solution was to delete the replica and to start over from that.

After a while I created a ticket and soon after that it turned out that I had not enough memory on the cdp proxy available (veeam one wasn't telling me that, but that's another story) during creation of the long-time retention point.

Of course it was my fault and I blame myself for not reading the specs properly enough, but having a corruption because of an aborted operation is probably not what the product was designed to do. Because any other failure like network interruption or a power outage might lead to the same consequences? Can you please review that to A confirm and B improve it in the future?

Thanks!

[HannesK]

Hello,
can you help me to find the support case number? Without case number with logs, it's unlikely to get any answer from QA. In general, we improve CDP in 12.1 significantly.

Best regards,
Hannes

[veremin]

The described behavior does not seem correct, so let us know if you have ticket numbers still available, so we can double-check the investigation. Thanks!

[mcz]

case # was 02933859.

[mcz]

maybe we have a bit a different understanding about the same words/situation but what I wannted to say was that the situation was unrecoverable. Once cdp has stopped working properly for one affected vm, I did not recover, even not in time (after some days). So I had to delete it so that it would become alive again. I would have expected it to recover automatically...

Very curious about 12.1

[veremin]

The R&D team will try to take a look at the previous investigation and confirm your findings. It will not happen immediately, as we are busy with the 12.1-related tasks, but I will surely update the topic, once we have more information regarding the case.

Thanks!