Proxmox node stuck during reboot – host does not shut down

[katti]

Hello Proxmox team,


I am experiencing a recurring issue on a Proxmox VE node that becomes stuck during reboot or shutdown and never completes the shutdown sequence.

Environment​

• Server: HPE ProLiant DL580 Gen10
• Hypervisor: Proxmox VE (clustered)
• Storage: Fibre Channel SAN (HPE 3PAR) using dm-multipath
• Cluster: 2 nodes + qdevice
• Watchdog: enabled
• iLO firmware: U34 (v3.60)

Issue description​

During reboot or shutdown, the node hangs indefinitely and displays the following messages on the console:

• the system does not power off,
• the reboot never completes,
• only a forced reboot via iLO allows the server to recover.

Observations​

• The issue appears during the shutdown phase, not during normal operation.
• SAN LUNs remain mapped during shutdown.
• The node is stable in production (no freezes or I/O issues).
• The watchdog is enabled but seems unable to stop correctly during shutdown.
• No explicit error is reported in the Proxmox web UI before reboot.

Checks performed​

• dm-multipath is correctly configured and operational.
• The Corosync cluster is healthy before shutdown.
• No running VMs at shutdown time.
• Watchdog is configured according to Proxmox recommendations.

Questions​

• Has anyone already encountered this behavior with dm-multipath and FC SAN storage?
• Are there any recommended kernel parameters or systemd overrides to ensure proper device-mapper cleanup during shutdown?
• Should the watchdog be disabled or tuned in Proxmox clusters using shared FC SAN storage?

Any advice or best practices would be greatly appreciated.


Thank you in advance for your help.