I’m proposing a new option to pbs_snapshot to quickly capture basic configurational/state data, as the full pbs_snapshot can take a long time depending on the site. This can be paired with other options like --accounting-logs or --daemon-logs in case users wanted to capture logs + basic configuration.
I think this idea is good, but I don’t like the name. You are collecting more than just config data. You are getting the jobs as well. How about --slim or something like that (I don’t like that name either). Instead of being overly paranoid like we usually are, we are just gathering the necessary subset to get an idea of what is going on.
@agrawalravi90 , @bhroam and anyone else reading this post.
pbs_snapshot in ‘Basic’ mode was not capturing subjob information from the system. To address this I’ve proposed a change to pbs_snapshot to capture output of qstat - tf command.
I’ve made the change in the design document. I request you to please have a look.
Thanks Ravi!
I thought of not removing qstat -f output because it is a published interface and if someone wants to just view the jobs without subjobs, this file would be helpful.
That said, since there is information being duplicated, I can remove it too, if you feel strongly about it.