I’m using core 20 on Raspberry Pi 4. Using the latest published snaps.
As soon as I set the watchdog runtime-timeout, systems starts rebooting every minute - regardless of the timeout. It appears the watchdog is being enabled, but never serviced.
Configuration before - as expected, runtime timeout not set, no 10-snapd-watchdog.conf file:
$ snap list
Name Version Rev Tracking Publisher Notes
core20 20210319 974 latest/stable canonical? base
pi 20-1 89 20/stable canonical? gadget
pi-kernel 5.4.0-1025.28 231 20/stable canonical? kernel
snapd 2.49.2 11584 latest/stable canonical? snapd
$ cat /etc/systemd/system.conf.d/10-snapd-watchdog.conf
cat: /etc/systemd/system.conf.d/10-snapd-watchdog.conf: No such file or directory
$ systemctl show | grep -i watchdog
RuntimeWatchdogUSec=0
RebootWatchdogUSec=10min
KExecWatchdogUSec=0
ServiceWatchdogs=yes
Next, I set runtime-timeout to 15 min. Resulting configuration looks good
$ snap set core watchdog.runtime-timeout=15m
$ cat /etc/systemd/system.conf.d/10-snapd-watchdog.conf
[Manager]
RuntimeWatchdogSec=900
$ systemctl show | grep -i watchdog
RuntimeWatchdogUSec=15min
RebootWatchdogUSec=10min
KExecWatchdogUSec=0
ServiceWatchdogs=yes
Approximately 60 seconds later, board will reset. It will boot successfully, run for about 60 secs, then reboot again. I can break it out of loop by clearing runtime-timeout.
Thoughts?
Dan