There should be a way to force a reboot, even if pods/mounts and other things hang.
Description
I had a ceph-rbd-mount that hung, which requires a reboot of the running node.
The reboot ended up taking more than 30 mins, ending with me physically rebooting the node instead of it being able to reboot.
The API was functional through-out the reboot-stage, so some button or command with similar functionallity as /proc/sysrq-trigger would be nice, the ability to configure a grace period (as mentioned in #4655 ) would also be nice.
Bug Report
There should be a way to force a reboot, even if pods/mounts and other things hang.
Description
I had a ceph-rbd-mount that hung, which requires a reboot of the running node. The reboot ended up taking more than 30 mins, ending with me physically rebooting the node instead of it being able to reboot. The API was functional through-out the reboot-stage, so some button or command with similar functionallity as
/proc/sysrq-trigger
would be nice, the ability to configure a grace period (as mentioned in #4655 ) would also be nice.Logs
Environment