[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Condor-users] Monitoring Condor
- Date: Sat, 9 May 2009 12:11:14 -0300
- From: Samir Cury <rimasy@xxxxxxxxx>
- Subject: Re: [Condor-users] Monitoring Condor
I don't get exactly what you need, but i think that hawkeye, that already comes with condor, can do part of the job.
I wrote a web interface for condor queue monitoring, that i'll announce soon here in the list, you can have a preview here:
i'm still working on the final version, but it can work for you as it is now, i think. If need any support can write me.
Although i think it won't help to monitor a GRID, hawkeye does more a resource monitoring on the nodes, it was what i understood, reading the documentation.
I heard about condorview too, but i don't recommend, as it seems was the doctoring thesis of a guy, and after that no more versions gone out, since 2006.
People at the cluster i work tried it some time but it drains too much resources (from machines and human efforts), for almost nothing interesting.
Ah, about your second question, if you tell us what is important to monitor in your GRID with Condor, everything will get easier to answer, but as a start, you can do a:
pick any job ID (the first column)
$condor_q -global -l <job_id>
So you can see all ads of your jobs, ads are like properties (can give you some idea of what monitor)
I think nodes have ads too (correct me if i'm wrong), so you can see their properties just doing :
pick any node
So you can see better what kind of information is interesting there
Hope that it helps.