Setting up HA: deleted nodes showing as master/LRM

[leccelecce]

I originally set up my proxmox cluster with 3 nodes: proxmox1, proxmox2, proxmox3. This cluster had high availability set up.

Later on I added some newer hardware to the same cluster, creating pmx4, pmx5 and pmx6. I then shut down the original proxmox1, proxmox2, proxmox3 and used pvecm delnode proxmox1, 2, 3 to remove them from the new ones..

Removing the old nodes seems to have worked everywhere in the cluster, *except* in the HA tab. For some reason, the original proxmox1 is still showing as the master, and nodes 1 and 2 are both showing as having the LRM running.




I did find /etc/pve/local/lrm_status but confusingly on proxmox4 for example this just contains the below, which isn't reflected in the UI anywhere.

{"mode":"active","state":"wait_for_agent_lock","results":{},"timestamp":1678223702}

I have had a dig through the /etc/pve directory but can't obviously see where this config exists. Can anyone suggest how I can reset the HA subsystem to get it working on the new nodes 4,5,6?

Thank you in advance for any support

 

[fiona]

Hi,
do you have any active HA services? Otherwise, the information here is not updated unfortunately. You can just add a HA service and wait for an hour or so until the old nodes are considered dead. Then the information here should be updated.

Do you have left-over directories for the deleted nodes in /etc/pve/nodes?

If you do have active HA services, please post the output of the following:

pveversion -v
ha-manager status -v
cat /etc/pve/ha/manager_status

 

[leccelecce]

Thank you Fiona, adding an active HA service solved it! I was nervous of doing that in case the old config caused problems but it's working fine. The old nodes vanished a short while agter.

Many thanks for your assistance.