Rolling Cluster Update Script

[thanegill]

I thought I would share a script that I created and have been using to do rolling updates for Proxmox clusters: https://github.com/thanegill/proxmox-upgrade-cluster

It doesn't cover all edge cases but does a decent job of waiting for the correct things to happen in the correct order. Issues and PRs are welcome; let me know what you think.

Features:

• Automatically determine cluster members (-c HOSTNAME).
• Upgrade all or specific nodes (-n HOSTNAME).
• Reboot nodes only when needed, or can optionally force a reboot (--force-reboot).
• Optionally force package reinstallation after upgrade (--pkg-reinstall PACKAGE).
• No-op mode (--dry-run).
• Multiple levels of verbosity (-v, -vv, -vvv).

 

[the_MiCrO]

Looks nice
Maybe you could add a few ceph specific commands.
At least: When rebooting a node with ceph disks set 'ceph osd set noout' and unset it after the node is back.