🔧 Maintenance Mode Stability Fix
Hotfix for maintenance mode dropping out immediately after activation on clusters with HA enabled.
Root cause: PVE removes the ha-manager maintenance flag within ~1 second when no HA-managed resources remain on the node. PegaProx's background polling detected this and removed the node from maintenance — even while VM evacuation was still in progress.
Fixes:
- Always evacuate VMs ourselves instead of relying on PVE HA to migrate them
- Set the HA flag first (so PVE doesn't fence the node), then actively migrate all VMs
- Background polling no longer removes nodes from maintenance that PegaProx put there
- Only auto-cleanup externally discovered maintenance states (e.g. set via Proxmox GUI)
📋 Updating
cd /opt/PegaProx
curl -O https://raw.githubusercontent.com/PegaProx/project-pegaprox/refs/heads/main/update.sh
chmod +x update.sh && sudo bash update.sh
Or over our Web updater.