Removing an Offline Node from a Proxmox VE Cluster

Affected Components

Proxmox VE (All versions with clustering)

Summary

A Proxmox VE node has gone permanently offline and must be removed from the cluster to maintain a healthy cluster state. Standard removal commands may leave behind stale entries, causing the node to persist in the UI with errors.

Symptoms

An unresponsive node is marked as “down” or has a red ‘X’ in the Proxmox web interface.
pvecm status shows the node as offline.
After attempting removal, the node may still appear in the UI, sometimes with a pve-ssl.pem' does not exist! (500) error.

Cause

When a node goes offline due to hardware failure or other issues, it must be formally evicted from the cluster. The standard pvecm delnode command may not always clean up all configuration entries, particularly in the underlying Corosync configuration, leading to a desynchronized state.

Diagnosis

Verify Node Status On a healthy cluster node, confirm the target node is offline.
Terminal window
```
pvecm nodes
```
Attempt Standard Removal Try the standard node removal command. If this completes without error and the node disappears from the UI, no further action is needed.
Terminal window
```
pvecm delnode <offline_node_name>
```

Resolution/Workaround

This procedure outlines the forceful removal of a node that is permanently offline.

Prerequisites:

The offline node must be powered down and will not come back online.
SSH and root access to a healthy node in the cluster.

Steps to Resolve:

Remove the Node from the Cluster From a healthy node, run the delnode command.
Terminal window
```
pvecm delnode <offline_node_name>
```
Clean Up the Node’s Directory The pvecm command may leave behind the node’s configuration directory. This must be removed manually to clean up the UI.
Terminal window
```
rm -rf /etc/pve/nodes/<offline_node_name>
```
Verify Removal Check that the node is no longer listed in the cluster’s node list.
Terminal window
```
pvecm nodes
```
At this point, also perform a hard refresh of the Proxmox web UI. If the node is gone, the process is complete. If it persists with an error, the Corosync configuration must be manually edited (Advanced Steps).

Advanced Steps (If Node Persists in UI):

Stop the Cluster Filesystem on the Local Node This is a prerequisite for manually editing the Corosync configuration.
Terminal window
```
systemctl stop pve-cluster
```
Edit the Corosync Configuration Open /etc/pve/corosync.conf with a text editor. Backup this file before editing.
Terminal window
```
cp /etc/pve/corosync.conf /etc/pve/corosync.conf.bak
nano /etc/pve/corosync.conf
```
- In the nodelist section, delete the entire node { ... } block corresponding to the offline node.
- In the quorum section, decrease the expected_votes value by 1.
Start the Cluster Filesystem Restart the service to apply the changes.
Terminal window
```
systemctl start pve-cluster
```

Important Considerations:

This procedure is for permanently offline nodes only. If the node might return, it should be properly shut down and brought back online instead.
Manually editing corosync.conf is dangerous. An incorrect edit can break the entire cluster.

References

Proxmox VE Wiki: Cluster Manager

Keywords

proxmox, pve, cluster, remove node, delnode, offline, corosync, pvecm

Disclaimer: The content of this knowledge base is compiled from a variety of publicly available sources, including community forums, mailing lists, blog posts, documentation, and other internet resources. While we strive to ensure accuracy and provide attribution wherever possible, we make no guarantees or warranties—express or implied—regarding the applicability, reliability, or effectiveness of the information in your specific environment. All credit goes to the original authors, and we have made efforts to include links or references to source material where available. If you believe any content requires correction, clarification, or updated attribution, please contact us here.