I performed the backup and restore of a PVE node that is part of a PVE cluster using the Proxmox Backup Client. This test assumes a scenario where a PVE node in the cluster experiences a failure.
1. Build a PVE cluster.
2. Backup the PVE node using Proxmox Backup Client:
3. Reinstall PVE on the node (assuming the node failure).
4. Restore the files using Proxmox Backup Client:
5. Synchronize the restored files:
6. Reboot the PVE node.
7. Delete
8. Add the recovered PVE node back into the cluster.
9. Confirm (presumably) that the recovery was successful.
---
I have the following questions regarding this procedure. I would appreciate your help.
- After following this procedure, I determined that recovery was successful because VMs/CTs were migrated back to the restored PVE node by `pvecm` and the HA feature. Are there any other points I should check to confirm that the recovery was done correctly?
- This method can restore the configuration, but it doesn’t restore the PVE version (such as the PVE kernel). While I can restore it by updating the PVE node, is there no way to restore the version directly from the backup?
*Note: I tried synchronizing the kernel with
- After syncing the configuration files using `rsync` and rebooting the system, the ServerID of the recovered PVE node was the same as at the time of the backup. In this case, does this mean I don’t need to reissue the Subscription key since the ServerID is the same?
- Is it only possible to restore files individually when restoring backups made with Proxmox Backup Client? Is there any way to restore them as an image?
1. Build a PVE cluster.
2. Backup the PVE node using Proxmox Backup Client:
proxmox-backup-client backup root.pxar:/ --repository 192.168.2.49:zfs3. Reinstall PVE on the node (assuming the node failure).
4. Restore the files using Proxmox Backup Client:
proxmox-backup-client restore host/pve-handon1/2024-09-14T19:05:09Z root.pxar /tmp/bk --repository 192.168.2.49:zfs5. Synchronize the restored files:
Bash:
rsync -av \
--exclude=/etc/pve/ \
--exclude=/etc/corosync/ \
/tmp/bk/etc/ /etc/6. Reboot the PVE node.
7. Delete
/etc/corosync/authkey8. Add the recovered PVE node back into the cluster.
9. Confirm (presumably) that the recovery was successful.
---
I have the following questions regarding this procedure. I would appreciate your help.
- After following this procedure, I determined that recovery was successful because VMs/CTs were migrated back to the restored PVE node by `pvecm` and the HA feature. Are there any other points I should check to confirm that the recovery was done correctly?
- This method can restore the configuration, but it doesn’t restore the PVE version (such as the PVE kernel). While I can restore it by updating the PVE node, is there no way to restore the version directly from the backup?
*Note: I tried synchronizing the kernel with
rsync -av /tmp/bk/boot/ /boot/, but the system failed to boot.*- After syncing the configuration files using `rsync` and rebooting the system, the ServerID of the recovered PVE node was the same as at the time of the backup. In this case, does this mean I don’t need to reissue the Subscription key since the ServerID is the same?
- Is it only possible to restore files individually when restoring backups made with Proxmox Backup Client? Is there any way to restore them as an image?