Proxmox Cluster Trouble Shooting
Restart Cluster Node Registration
systemctl stop corosync.service
systemctl stop pve-cluster.service
systemctl stop pvedaemon.service
systemctl stop pveproxy.service
systemctl start corosync.service
systemctl start pve-cluster.service
systemctl start pvedaemon.service
systemctl start pveproxy.service
TCP Dump Multicast Cluster Traffic
tcpdump -n -i eth1 multicast | grep 5404
Print (to console) Corosync Log
journalctl | grep corosync
Print corosync Service Status
systemctl status corosync.service
Force Quorum
Edit /etc/pve when cluster is down
systemctl stop pve-cluster
pmxcfs -l
killall mcxcfs
systemctl start pve-cluster
A note about Multicast Traffic
corosync and Proxmox clustering need multicast traffic to be configured properly for cluster nodes to see each other.
If a node keeps losing contact with other members make sure IGMP snooping is turned off at the switch and layer 2 level.
If some nodes connect initially then disconnect after a few minutes take a look at the multicast configuration of the switches these nodes connect through.