[GE users] Memory hard limit seems to be ignored on sge 5.3p5 version

Rayson Ho raysonho at eseenet.com
Mon Sep 13 18:16:24 BST 2004


    [ 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. ]

1) What OS are you running??

2) The shepherd should set the process limit before starting the job. You
can check the soft and hard limit inside the job by putting "limit" and
"limit -h" (for tcsh), or "ulimit" and "ulimit -a" (for bash, ksh, ksh) in
the job script. Then send us the output.

Rayson


>Hi,
>
>   I have  a cluster with sge 5.3p5 installed.  I set the memory hard
limit to 1.5 GB for the queue through h_data limit.  But the job seems to
be using 1.9 GB as per top command and is still going ok. Also output of
qhost shows 1.9 GB MEMUSE as well. I was expecting SGE to kill the job once
it exceeded 1.5 GB.   Is there any way to enforce the limit of memory usage
to 1.5 GB?  Thanks.
>
>Prakashan
>
---------------------------------------------------------
Get your FREE E-mail account at http://www.eseenet.com !

---------------------------------------------------------------------
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