Orphaned Node in GUI und Corosync

[Henrik Gießel]

Hello,

I have problem in one of our clusters. The cluster had 2 Node. Yesterday I wanted to add a 3rd one. Due to a mistake in the IP-Config the Clusterjoin weren't finished propably. I decided to reinstall the 3rd node with new IP and FQDN. This time it works fine. But now I have an oprhaned node shown in the Web-UI:

- I already deleted /etc/pve/nodes/<FAILED NODENAME>
- with "pvecm status" I have 4 exptected votes and currently 3 votes
- "pvecm nodes" dont show the failed node. I think because the join-process doesn't finished successfully
- /etc/pve/corosync still has a entry with the IP of the failed Node
- The WEB-UI still shows the orphaned node

- Some operations in the WebUI cant be done. For example when I want to migrate a VM I need to define a target Hosts. The Dropdown-menu for this doesnt show any entries. Migratating and other Tasks are still possible over the CLI.

How can I aproach this issues? Can I delete the entry of for the failed node manually from corosync and restart the corosync service and pveproxy? What else I have to do?

PS: I use PVE 5.3 and higher

 

[mira]

Yes, you should be fine deleting it from the corosync config. You might want to bump the version as well as explained here: https://pve.proxmox.com/pve-docs/chapter-pvecm.html#_corosync_configuration

 

[Henrik Gießel]

Yesterday I tried it. It works. Ty for your help