TL;DR Do away with the High Availability stack in a safe manner. Avoid the risk of reboots when troubleshooting suspected hardware woes. Particularly useful for frequent loss-of-quorum scenarios.
ORIGINAL POST How to disable HA permanently
This only helps you, obviously, in case you are NOT using HA. It is also a sure way of avoiding any bugs present in HA logic which you may otherwise encounter even when not using it. It further saves you some of the wasteful block layer writes associated with HA state sharing across nodes.
TIP The rationale for this snippet is covered in a separate post on High Availability related watchdog that Proxmox employ on every single node at all times.
All CLI examples tested with PVE 8.2.
You have to stop the HA CRM & LRM services first, then the multiplexer, then unload the kernel module:
systemctl stop pve-ha-crm pve-ha-lrm systemctl stop watchdog-mux rmmod softdog
To make this reliably persistent following reboots and updates:
systemctl mask pve-ha-crm pve-ha-lrm watchdog-mux cat > /etc/modprobe.d/softdog-deny.conf << EOF blacklist softdog install softdog /bin/false EOF
TIP There is a separate snippet if you are only looking to disable HA temporarily, such as for maintenance.