[GE users] qacct too slow

Jinal Jhaveri jajhaveri at lbl.gov
Tue Mar 28 01:20:34 BST 2006


    [ The following text is in the "ISO-8859-1" character set. ]
    [ Your display is set for the "ISO-8859-10" character set.  ]
    [ Some special characters may be displayed incorrectly. ]

I am trying to check status of a SGE  job started via web and found that 
qacct is too slow due to the huge size of accounting file. I was 
wondering how to avoid this delay. Several ideas I can think of:

1) Do a log rotate on the "accounting" file to avoid it being too big.

2) Write logs to a database/ another flat file on a queue basis (with 
log rotate) using the epilog . The problem with this approach is that 
there are some types of errors which might happen even before the job 
starts. (For e.g the directory from which the job is running doesn't 
exist on the execution node). I think that these types of errors happen 
before prolog/epilog and are reported to the qmaster. Does anybody know 
how to track these errors? Is there any execd parameter I can set that 
will allow me to write this error logs to a file/database using the 
prolog script/ anything else?



Any suggestions on this would be highly appreciated. What are my other 
options? I am using SGE 6u6


thank you
--Jinal

---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
For additional commands, e-mail: users-help at gridengine.sunsource.net




More information about the gridengine-users mailing list