This message was deleted.
# longhorn-storage
a
This message was deleted.
a
I've had all sorts of strange issues with Longhorn. The most recent is when I tried upgrading (via Rancher UI) from 1.5.3 to 1.6.2. It looked like the upgrade went fine, but I couldn't attach any volumes the next day. The v1 data engine box wasn't checked by default. Every time I checked the box in the settings, the next day it was reverted. So I tried going back into Rancher Apps to edit the config and redeploy, but it failed saying that Longhorn was still upgrading. This was like a week later with everything indicating the Longhorn was at 1.6.2 and I did not see any reason why it would still be updating. I have not figured it out yet. Luckily this is just our test cluster, if that happened in prod I'd have hundreds of angry users. Sorry, wish I could help, we put off any Longhorn upgrades exactly for these reasons
f
It's all good
I think the crux of my issue was that I "upgraded" from the longhorn official chart (v1.4.0) to the Rancher longhorn chart (103.1.1+up1.4.4). The Rancher longhorn chart doesn't include CRDs, and expects you to have a separate chart installed (longhorn-crds).
So at least I was able to identify the issue
helm log from the "upgrade"