[Question] Who will update the NodeCondition to "False" after the problem recovery

[Congrool]

Hi, I'm using the system-log-monitor to detect the node problem and update the respective NodeCondition, which can match logs in /dev/kmsg of specific pattern.

I noticed that the NPD updates the NodeCondtion Status to True when it gets the target log. But after digging into the code of systemlogmonitor, I didn't find anything which will reset the NodeCondition to False after recovery.

I want to know who will take the responsibility to update the NodeCondtion to False after recovery in the best practice. Is it the job of remedy system?

[severloh]

I also couldn't find this feature. Currently, my solution is to write a program to reset the status of the nodes manually.

[googs1025]

IIUC, this needs to be done manually. We currently use a custom script to implement this internally. npd is a inspect system that is responsible for reporting problems but not for recovery. Recovery requires manual recovery and manual recovery of the status.