[GE users] jobs pending

pvdmeer pvdmeer at gmail.com
Thu May 14 19:28:51 BST 2009


    [ 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. ]

Hi Chris,

Thanks for the advice. I finally got stuff running. "qstat -f" is quite essential indeed.

What I did was just reboot my main Ubuntu box. And then I perform the following:

> sudo apt-get purge gridengine*
> sudo apt-get install gridengine*

(enter default cell, <hostname>)

> sudo qmon

("Queue Control" menu -> add queue 'all.q' with <hostname> in it)
("Host Configuration" menu -> "Submit Host" tab -> add <hostname>)

> cd <examples/jobs directory>

> ./pascal.sh 3

(it works but it ain't fast)

I'll add my other boxes to the cluster now as well.

With kind regards,

Pieter


On Thu, May 14, 2009 at 12:35 PM, craffi <dag at sonsorol.org<mailto:dag at sonsorol.org>> wrote:
Does the output of "qstat -f" show a queue instance in state "au"? If
so that means that the sge_execd is not running.

If you see no queues than it is likely that the exechost functions did
not get installed at all. Look in the SGE_ROOT directory for a script
called "inst_sge" - you can call that as root via "./inst_sge -x" to
try reinstalling the execution host component.

Chris



On May 14, 2009, at 4:26 AM, pvdmeer wrote:

> Dear Grid Engine users,
>
> Recently I installed Sun Grid Engine on my own cluster of Ubuntu
> machines. Previously, with a manual installation under Slackware, I
> was able to run the example scripts (like pascal.sh). But with my
> new OS, at this moment, I seem to be stuck. The installation using
> the apt-get package manager is really a lot simpler. But I may have
> missed some crucial configuration steps..
>
> My problem is the following. I try to submit a job. This succeeds
> but when using qstat the job remains pending for infinity. It shows
> the status 'qw' or 'hqw'. Also when doing a 'ps aux | grep sge':
> sge_qmaster shows up but I cannot find sge_execd. Somehow, I have
> the feeling my machines are not configured to be execution nodes.
> And trying to run sge_execd has no effect. There should be some way
> to get that daemon running??
>
> Has anyone encountered this problem before? I'd be grateful for any
> help. Thanks in advance.
>
> With kind regards,
>
>
> Pieter van der Meer
>

------------------------------------------------------
http://gridengine.sunsource.net/ds/viewMessage.do?dsForumId=38&dsMessageId=195460

To unsubscribe from this discussion, e-mail: [users-unsubscribe at gridengine.sunsource.net<mailto:users-unsubscribe at gridengine.sunsource.net>].




More information about the gridengine-users mailing list