This message was deleted.

I have had this issue several times, caused by some trailing label or annotation. I can’t remember which, but I’m sure you’ll recognise it when you read them all.

Could you please file an issue to harvester? https://github.com/harvester/harvester/issues Providing a support bundle would also help in gaining a better understanding of what's going on.

The reason was that the disk-ID was different - we had reinstalled two of the nodes but kept the same name- somewhere in Harvester the old disk ID was saved so it screwed up Longhorn. It took some time to find this but after reinstalling the nodes with random name instead the problem was solved. Could be good to specify in the documentation that it is not recommended to reuse on old name for a node- or specify a way to remove the old disk id (we tried to change directly in the yaml files- it didnt work)