Is there a way to use a /dev/watchdog ?
Hi,
I run a ceph cluster on docker on a linode vps.
Sometimes the cluster fails and the vps keeps trying to reach the failed ceph disk and wont reboot.
Sending something to /dev/watchdog, does not reset the machine.
Is this a known issue? Or should i configure something in the linode manager to get this working?
Pushing "Rebooot my machine" on the webstite works just fine.
Kind regards,
Ed
2 Replies
jyoo
Linode Staff
You should be able to configure a software watchdog. There is some information available here regarding the use of modprobe to initiate that. Alternatively, you can do a command line reboot as long as you have the Lassie watchdog enabled. Lassie will sense that your Linode was "unexpectedly" shut down and can reboot it for you from there. Another option would be to use the Linode API, particularly if you wanted to set up a script to reboot your Linode automatically.
Thanks, i ended up doing just that, using the api to reboot the machine.
RAAS ( Reboot As A Service ) ;-)