We have a file system that is not always available, but would like to have it in the OOD Files menu when it is. Is there anyway to set a timeout for the FavoritePath.new so it continues processing w/o that FS if it doesn’t return in a specified time? Currently if a user tries to log in and the FS is not responding, OOD will hang until it responds before continuing.
This might be something better done in another check before adding the new path, but I’m still a Ruby neophyte.
Thanks,
Bruce
This is the best option that we’ve come up with. We have prometheus alerting about filesystem outages . Which failed us last outage - the FS was showing as available when it wasn’t. So we added the additional check on a file /etc/ood/config/gpfs_outage that we can manaully drop if we know there’s an outage but for whatever reason Prometheus doesn’t report it.
Ruby’s own Timeout library doesn’t really help or work well and it’s generally advisable to not use it.
We’ve found that basically any command to reach the FS is going to hang. The Prometheus check itself is on crontab which seems to have different behaviour. FS checks in a crontab don’t seem to hang, so this strategy of checking in a crontab and creating or populating a file that OOD reads seems to work OK.
Jeff, this sounds like a good solution for us as well, or at least something very similar.
Thanks,
Bruce
This topic was automatically closed 180 days after the last reply. New replies are no longer allowed.