[gridengine users] funny output order

Reuti reuti at staff.uni-marburg.de
Thu Jul 18 19:13:56 UTC 2013


Hi,

Am 18.07.2013 um 17:14 schrieb Ulrich Hiller:

> i have a funny problem where i do not have an idea what is going on.
> 
> First: we have an opensuse 12.3 cluster with GE2011.11p1
> The master node has the /home and all the nodes have /home nfs monted.
> Anything else (/ /var /tmp) is locally on each node.
> 
> Now i have a script (on /home/username - local on master, nfs mount on
> nodes):
> #!/bin/bash
> #$ -S /bin/bash
> #$ -N IDLname
> #$ -V
> #$ -p -10
> #$ -m n
> #$ -j n
> #$ -l h_rt=12:00:00
> #$ -t 1-158

Is it possible, that several instances of this array job are running at the same time no one and the same host?

I see only 2 `grep` in your job script but 4 in the output. It would be good to use an unique one in $TMPDIR (okay, will be deleted after job if not saved).

FWIW: Instead of redirecting the output of each statement, the bash can also provide a redirection like:

exec 3>&1
exec 1>$TMPDIR/i
echo 111
echo 222
exec 1>&3
exec 3>&-

-- Reuti

> echo "+++++++++++++++++++++++++"  > /tmp/i
> hostname >> /tmp/i
> echo "11111111111111111111" >> /tmp/i
> ls -la /tmp/.X*-lock >> /tmp/i
> cat /tmp/.X*-lock >> /tmp/i
> echo "2222222222222222222" >> /tmp/i
> ps -ef | grep -i xvfb >> /tmp/i
> echo "--------------------"  >> /tmp/i
> echo "33333333333333333" >> /tmp/i
> ls -la /tmp/.X*-lock >> /tmp/i
> echo "44444444444444444444" >> /tmp/i
> ps -ef | grep -i xvfb >> /tmp/i
> echo "........................................." >> /tmp/i
> exit
> 
> 
> Which i run:
> qsub -cwd -o output.dat -e error.dat test1.sh
> 
> The /tmp/i (local on each node) looks e.g. like this:
> +++++++++++++++++++++++++
> bee003
> 2222222222222222222
> 11111111111111111111
> linadm    9952  9804  0 16:14 ?        00:00:00 grep -i xvfb
> .........................................
> 2222222222222222222
> linadm    9957  9856  0 16:14 ?        00:00:00 grep -i xvfb
> --------------------
> 33333333333333333
> 44444444444444444444
> linadm    9961  9878  0 16:14 ?        00:00:00 grep -i xvfb
> --------------------
> 33333333333333333
> 44444444444444444444
> linadm    9964  9856  0 16:14 ?        00:00:00 grep -i xvfb
> .........................................
> linadm    9967  9878  0 16:14 ?        00:00:00 grep -i xvfb
> 
> 
> Different runs give different output orders (seems to be random).
> 
> Does anybody have an idea what is going on there? Why doesn't it run the
> script line by line?
> 
> Any help is appreciated.
> 
> With kind regards, ulrich
> 
> 
> -- 
> ------
> Ulrich Hiller
> Max-Planck-Institute for Astronomy
> Koenigstuhl 17
> 69117 Heidelberg
> Germany
> phone +49 6221 528238
> fax +49 6221 528246
> e-mail hiller at mpia.de
> 
> _______________________________________________
> users mailing list
> users at gridengine.org
> https://gridengine.org/mailman/listinfo/users





More information about the users mailing list