finally we did our 13.1 (from 12.1) upgrade. Mostly it worked well (still on Windows/MSSQL) and upgrade was quite fast but we had some issues:
Proxmox Backup did no longer work after upgrade (Case 08206449): Our self-signed certificate was from 2019 which lead to "Waiting for available workers" and workers unable to connect. At first the SSL message was not even visible and it looked more like an idrac interface on Veeam Server is causing the issue. Renewing the Veeam server certificate fixes the issue. Sadly, Veeam support was not very helpful and also only saw the interface at fault.
Component update with warnings (Case 08206359): One of our Linux repos showed warnings in the component update. We do not know what the warning was and also support could not tell us. The Linux server in question was in healthy state, without any upgraded needed directly after the warning.
Two of our VMs normally using persistent guest agent failed as connection could not be established via this agent - using non-persistent agent works right away - we might open another case for that.
VSphere Tag issues (Case 08207269): All our jobs use exclusion for one VSphere Tag Category. Since yesterday, some of the jobs "fail" (the VMs are still backed up but it fails nevertheless) because the tag category is "unavailable" (which never happened with 12.1 and even with 13.1 is transient). I that is true we might have a larger issue as some jobs are configured to backup everything in a folder "but" the tag category which is backed up by another job.
All in all this is one of the smoothest major version upgrade!
We searched our VSphere logs and found that according to Envoy logs VBR calls "DELETE /rest/com/vmware/cis/session" (deletes its VCenter login session?) and then "GET /tagging/tag" for which it at first gets a 401, after witch it reauthenticates.
This might be connected as between logout/reauthentication there is nearly a minute of time. Maybe this is all a session race between parallel workers?
Thanks for sharing your upgrade experience, and glad to hear that aside from the listed issues everything was mostly smooth
For the tag issue, please do continue with Support; your investigation on the vSphere logging side is helpful, and I can see the Engineer is currently reviewing the most recent update from your side.
I'm not sure if your case and olafurh's have the same root cause (at first blush error stacks I can see in the case notes look different), but as the investigations are both on-going, let's see what the Support Engineers end up finding.
David Domask | Product Management: Principal Analyst
