How to work with NodeHealthCheck.py (unsupported)

Users/Site Administrators
1

Hello good Morning.

Has anyone managed to use the NodeHealthCheck.py hook that comes with OpenPBS (under unssuported dir)?

I imported .py and .json as it comes and made a submission to a umonted /home node, but it is not taking the node offline. I have also changed the settings in json file from Warn to Offline, but no success.

The qmgr -c ‘l h’ says that the hook is enable:

Hook NHC
type = site
enabled = true
event = execjob_begin,execjob_prologue
user = pbsadmin
alarm = 30
order = 1
debug = false
fail_action = offline_vnodes

Any advice?

Thanks

2

You might have to create your own health check script that returns 0 (success, no issues) and 1 (issues, problems with the node). Call this to the node health check script from within this hook to set the nodes offline , free etc with a comment.