This message was deleted.
# harvester
a
This message was deleted.
r
Hi @sparse-jelly-7856 would you mind generating a support bundle and sharing it here so that we can look into it?
There might be some old resources for previous upgrades that need to be clean up
s
There were, I ad to kubectl delete managedchart -n fleet-local hvst-upgrade-xxxxx-upgradelog-operator and then go through and clean some other things up. Once that was done, I started the download again and it was taking a while so I went into the cluster group and force updated and then things started to work after that.
I finished the upgrade yesterday, so I'm not sure if the bundle will do anything but I can give you what I have now if you want it still?
c
Did you have a leftover helm deployment you didn't uninstall first? Sometimes even when running a failed helm install an installation is left around that needs to be uninstalled first. If not then it sounds like helm didn't clean some stuff up that it should have, which is confusing. If that was the leftover helm installation from the initial installation then you might have just needed to pass in the --upgrade command line option? Just guessing here, but they are educated guesses.
s
I'm not sure about the helm installatioin. I've never ran anything custom on this cluster and I'm not the most familiar with harvester. I ran the pre-check upgrade command and and they came back with all passes. The only addon I enabled was the rancher monitoring and the pci devices addon. I am totally guessing here also, and mine are completely uneducated guesses, but I think it didn't clean some stuff up from the last upgrade. The hvst-upgrade-xxxx did correspond to the previous upgrade I did to 1.2.1 a while ago. Or so I think it was the same, I can't be for sure, all i can remember from that upgrade is that when it was setting things up and I happened to glance at longhorn I saw a volume that had the hvst-upgrade-x and the x was the same number.
r
Did the upgrade succeed after you removed the ManagedChart? The previous upgrade Logging and ManagedChart objects remaining in the cluster seem to prevent the new upgrade from being started. Alternatively, you can initiate the upgrade with the “Enable upgrade logging” unchecked to fully disable the upgrade logging feature.
s
Yes as soon as i removed the chart I was able to reinitiate the upgrade
👍 1