Node rejected from cluster since IP change.

[tprice]

Hi all,

I changed the IP of a node in a cluster, first I updated /etc/network/interfaces, /etc/host and corosync.conf while incrementing the version.

I then removed the cert in /etc/pve/pve-root-ca.pem and key in /etc/pve/priv/pve-root-ca.key on the node with the changed IP. I also removed pve-ssl.key and pve-ssl.pem from the updated node in /etc/pve/nodes/updatednode/. Then I ran pvecm updatecerts -f and can now see the new IP updated on the cert in the webgui of the updated node.

I then went to each node in the cluster and ran pvecm updatecerts -f on each node in the cluster. Im still not seeing pve-ssl.key and pve-ssl.pem in /etc/pve/updatednode/ from another node that wasn't updated, can see the files on the updated node. I suspect this is the issue but I do not know why the files are not being added for the other nodes? Maybe I have to manually move the pve-ssl.pem an key to each node as the updated node is currently isolated? I will try this another day but was hoping for some insight in the meantime.

 

[leesteken]

Since it is a cluster, why not remove the node (with its original IP), wipe it, reinstall Proxmox and add it as a new node (with a new name and IP)? That's what clusters are designed for: replace nodes not tinker with them.

 

[tprice]

Since it is a cluster, why not remove the node (with its original IP), wipe it, reinstall Proxmox and add it as a new node (with a new name and IP)? That's what clusters are designed for: replace nodes not tinker with them.

The issue I'm faced with is the node needing its IP changed is in a different subnet than the rest of the nodes with a large amount of VM data that needs to be moved, without taking the VMs down before being wiped. If I can change the IP I can use a 10Gbps connection. Otherwise I'm forced through a slow firewall.