[GE users] SGE questions

Reuti reuti at staff.uni-marburg.de
Thu Mar 1 18:05:35 GMT 2007


Am 01.03.2007 um 18:29 schrieb john.li at mindspeed.com:

>
> Hi Rayson,
>
> Thanks for the reply.   And thanks for correct my mistake on the  
> description.
>
> The reason that I brought this behavior up is due to the umask  
> issue that I mentioned
> in my original email.   The umask seems to inherit user shell umask  
> when a job is
> submitted as a script.  (Inherit environment umask is what I  
> want)   but not as a binary.
> But SGE doesn't use $PATH to search for a script, which is not good  
> as far as I concern.
> I don't have this problem in LSF at all.

You mean something like:

qsub `which myscript.sh`

Is LSF spooling the jobscript, or using the original during the later  
execution?

-- Reuti


> Thanks,
>
>
>
>
> "Rayson Ho" <rayrayson at gmail.com>
> 02/28/07 03:07 PM
> Please respond to
> users at gridengine.sunsource.net
>
>
> To
> users at gridengine.sunsource.net
> cc
> Subject
> Re: [GE users] SGE questions
>
>
>
>
>
> On 2/28/07, john.li at mindspeed.com <john.li at mindspeed.com> wrote:
> > I'm new to SGE and trying to deploy SGE to replace LSF.
>
> Good :)
>
> > 3.  It seems that if I use -b y in the qsub  command, I MUST  
> specify the
> > full path
> > for the binary executable or it is in the current working  
> directory.   SGE
> > doesn't
> > use the $PATH variable to locate the binary executable.   But SGE  
> will use
> > $PATH
> > to locate a script if I use -b n for a script.   Is there a  
> reason for this
> > behavior?
>
> Shouldn't it be the other way around??
>
> "-b n" tells qsub that the job is a script, and SGE will look for the
> job script in the current directory or the location specified at qsub
> time.
>
> "-b y" tells qsub that the job is a binary program, and SGE will not
> spool the binary to the master machine. Since the location of the
> binary can be different with different OSes/architectures, recording
> the path does not make sense.
>
> I've just tested on my cluster and that's the behavior... and I
> believe it has something to do with POSIX 1003.2d (standard for batch
> queuing systems).
>
> * also note that LSF does not spool the job script or the binary at  
> bsub time.
>
> Rayson
>
>
>
>
> >
> > I have more questions, but this is good for now.  :-)     
> Someone,  please
> > help...
> >
> > Best regards,
> >
> > John,
> >
> >
> >
> >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
> For additional commands, e-mail: users-help at gridengine.sunsource.net
>
>

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