- set CT to "ignore" in the HA web GUI
- delete CT
service 'ct:123' without node
- Thread starter grin
- Start date
-
- Tags
- bug ha cluster
It seems to be easily fixed by manually cutting the respective entries from /etc/pve/ha/resources.cfg, with no visible side effects. Still, not ideal.
To check configuration consistency in all cases is a question of implementation effort. Since no harm is done in that case such checks are not in focus.
While I see your point in this case Proxmox deletes a container with active HA state, and HA subsystem is unable to handle this (cannot delete it). The problem is not [only] that there is no consistency check when deleting something still in HA but that along with this missing check there is a check in HA when trying to remove it (and realise that the CT isn't available anywhere) and it gives up.
Obviosuly there is a specific group of people which is able to handle this (and risk screwing it up since there is no guarantee that manually fiddling with config files will leave the system in a consistent state), but I guess the majority of people cannot fix this. It is always your choice to ignore a bugreport or fix it up (in this case it's probably an oneliner fix).
My problem is resolved… <roll_eyes />
I get flooded with these log entries, too.
Is there no better way than manually editing resources.cfg?
Is there no better way than manually editing resources.cfg?