Hello we had a very serious problem with the Proxmox cluster on OVH, our vRack stopped working.
We tried to solve the situation by hot adding a new corosync configuration with a redundancy link in it.
The configuration worked and corosync saw all members, we then restarted the proxmox cluster on each node, but it still saw the old configuration on the pmxcfs, we then changed the contents of the pmxcfs on each node by mounting it locally with the pmcfs -l command.
We then restarted systemctl restart cororosync and pve-cluster on all nodes. the services restarted but the pve-cluster sees a different Ring-ID on each node, and the cluster do not see other nodes.
What can we do to fix this?
Is pmxcfs unrecoverable after local mount ?
We have now been forced to put all nodes with pmxcfs in local mount to run backups.
We tried to solve the situation by hot adding a new corosync configuration with a redundancy link in it.
The configuration worked and corosync saw all members, we then restarted the proxmox cluster on each node, but it still saw the old configuration on the pmxcfs, we then changed the contents of the pmxcfs on each node by mounting it locally with the pmcfs -l command.
We then restarted systemctl restart cororosync and pve-cluster on all nodes. the services restarted but the pve-cluster sees a different Ring-ID on each node, and the cluster do not see other nodes.
What can we do to fix this?
Is pmxcfs unrecoverable after local mount ?
We have now been forced to put all nodes with pmxcfs in local mount to run backups.