PVE8.2, how to change node IP? Stuck on corosync propagation

[Rxunique]

I follow this old thread https://forum.proxmox.com/threads/change-ip-of-cluster-node.106676/#post-459672 and https://pve.proxmox.com/wiki/Cluster_Manager trying to change IP of my 3 nodes cluster.

Previously they were setup with no vLAN consideration, now I'm setting vlan meaning new subnets.

I've figured out that there're now 2 corosync.conf files, and how to edit both of them (previous post seems to be pre PVE7.2), I've also broke my cluster and figured out that manually reseting both corosync.conf along with other files can restore connection.

Here's overall my steps on node1

systemctl stop corosync pve-cluster && pmxcfs -l

nano /etc/pve/corosync.conf

nano /etc/hosts

nano /etc/network/interfaces

nano /etc/pve/priv/known_hosts

nano /etc/corosync/corosync.conf

killall pmxcfs && systemctl start pve-cluster corosync 

systemctl restart networking.service

I've tried to all 4 combination on node2 but just couldn't the cluster to work with new node1 ip

• with and withoutmodify /etc/corosync/corosync.conf
• with and without restarting corosync
  

Has anyone successfuly changed IP on pve8.2?

 

[Rxunique]

I got it to work after lots lots tinkering, but couldn't understand why.

First I made sure all nodes are updated to same micro versions, 8.2.7, not that this would be the cause, just to remove one unlikely variable.

Here's my exact sequence

0) starting points is all 3 nodes with old ip in quorum, config version3

1)updated node1 to config V4 with new ip, but without restarting network. Result is node1 separated from cluster, still accessible from old ip. Node2&3 in sync and in quorum but won't sync V4 config.

2) Restarted network on node1 to activate new IP, node2&3 still won't fetch V4. Tried on all nodes reboot, restart corosync etc etc.... they just won't sync V4

3) manually changed node 2 /etc/corosync/corosync.conf to V4, tried on all nodes restart corosync etc, not helping

4)manually stopped node2 corosync pve-cluster and changed /etc/pve/corosync.conf to V4, tried on all nodes restart corosync. Then it worked in one of my attempt when I noticed systemctl restart corysync took too long to respond, node3 also got V4 config.

Can guru here help shed some light on the underlying cause? Or what's should be the correct steps

 

[Rxunique]

Furthermore, I tested to change only /etc/pve/corosync.conf on node1 and restart pve-cluster, this will separate node1 from cluster immediately

so my current best theory (to be confirmed), is that this will brake node1 off the cluster, remaining 2 nodes on the cluster still have quorum and old config, so they won't sync new version from node1.

But somehow changing /etc/pve/corosync.conf on node2 won't break cluster & quorum so the new version can propagate from there.

 

[Rxunique]

I got all 3 nodes moved to new IP, when I changed Node1, and corosync of node 2, I was able to get node 3 to sync off from node 2

but when I changed node2, and node 3 IP one by one, couldn't get any remaining node to sync at all, end up manually changing corosync every node every time

 

[fiona]

Hi,
please try following the guide here: https://forum.proxmox.com/threads/corosync-cluster-problems-help.55856/#post-257275