[torqueusers] How to debug torque/maui

[Jacques Foury]

Albert Shih a écrit :
> Hi all,
>
>
> Stupid question from a newbie.
>
> I'm suppose to manage a cluster torque/maui, but when the are some problem
> I don't known the tools (like grep/vi) and where to find the problem.
>
> For example : 
>
> 	Actually I've a user submit through qsub a job, this job don't
> 	start. I think I known the reason. But how/where can I find some
> 	information about this job ? 
>
> where can I find some document like «torque/maui for dummies-sys-admin» 
>
> Regards.
>   

Hi Albert.

The main log files are :
/var/spool/maui/log/maui.log

you can specify its size in maui.cfg :

LOGFILE                 maui.log
LOGFILEMAXSIZE          50000000
LOGLEVEL                3
LOGFILEROLLDEPTH        3


For Torque it's not easier :

/var/lib/torque/server_log/20070830   (one file each day)


If you know the node where the job goes (if it starts), you have another
log file there in /var/lib/torque/mom_log

Then you become a cop : look for the events in each file to understand
what happens ;-)


And about the tools, I use grep, vi, awk, cut and so on ;-)

-- 

Jacques Foury
Institut de Mathématiques de Bordeaux
Université Bordeaux 1 / CNRS