Watchdawg

Proxmox VMs

I got tired of interrupting calls from the trouble and strife (and the rest of the family) that something wasn’t working. Seems to normally be weird edge cases between hardware and software and virtualization, but at least once the virtualized Ubuntu 20.04.4 running Plex has died and the OPNsense router died once too. So…. watchdogd!

On Proxmox I enabled the PCI watchdog device by adding:

watchdog: model=i6300esb,action=reset

to the end of the QEMU config file. Next in the Ubuntu host I installed watchdogd:

apt install watchdog

Updated /etc/watchdog.conf to include:

watchdog-device = /dev/watchdog
log-dir =  /var/log/watchdog
realtime = yes
priority = 1

And in /etc/default/watchdog:

watchdog_module="i6300esb"

Shutdown the system (not rebooted) and after powering back on I can see:

$ lspci -nn | grep -i watch
06:04.0 System peripheral [0880]: Intel Corporation 6300ESB Watchdog Timer [8086:25ab]

I wanted to do the same thing on an OPNsense VM. After configuring Proxmox the same way I added ichwd_load="YES" to /boot/loader.conf (can be done by adding it to System -> Settings -> Tuneables too). Same process of shutting down completely. After coming back up:

# pciconf -l -v | grep -i watch
    device     = '6300ESB Watchdog Timer'

Hardware

Lastly I wanted to set this up on the SMC-based router I use. Looking at the manual I saw:

So that NCT6776F is gonna be from the wbwd.ko. This time I added it to System -> Settings -> Tuneables in the GUI. wbwd_load with a value of YES. I rebooted, enabled watchdog in the BIOS and… Reboots. Didn’t work. Poking around I found watchdog_enable=yes should be added to /etc/rc.conf? I tried adding to tunables and that didn’t work. Once I edited /etc/rc.conf and rebooted everything worked. Hopefully there are auto-reboots in my future.

Leave a Reply

Your email address will not be published. Required fields are marked *