[SOLVED] I dread every single time I have to touch any thing related to corosync

[ishanjain]

It looks like pve-cluster service is responsible for it. I tried the steps here and still no changes are made to /etc/pve/.members

https://blog.dob.sk/2018/07/09/fix-invalid-node-ip-in-etc-pve-members/

 

[waltar]

Like this post it looks possible by hosts ... https://forum.proxmox.com/threads/edit-members-file.127584/

 

[waltar]

An IP change isn't intended by the pve developers ...

 

[ishanjain]

Okay, It's fixed! and for future reference, The right way to fix this problem is like this.

I'll take two nodes in my example. pvenas(10.0.99.9), pvepc(10.0.99.20 changed to 10.0.99.4)


1. Update /etc/hosts on pvepc by adding an entry like

10.0.99.4 pvepc.home.arpa pvepc

2. Run systemctl restart pve-cluster

3. Update config_version in /etc/pve/corosync.conf for good measure and then run systemctl restart corosync



I was editing /etc/hosts and restarting pve-cluster/corosync on pvenas machine instead of pvepc.

 

[ishanjain]

Interestingly, After all of this /etc/issue remains unchanged but it's a more aesthetic thing so I don't care about it.

 

[waltar]

/etc/issue is normal file which could be manually updated too, it's just filled by pve installation.