[torqueusers] qsub patch for job script arguments.
danielfcoimbra at gmail.com
Mon Aug 29 13:00:32 MDT 2011
I believe a reasonable alternative would be to be able to repeat the
option for every argument, as follows:
qsub -F arg1 -F arg2 -N test -F arg3 run.sh -F arg4
In this way, each occurrence could be considered a separate argument as if:
run.sh "arg1" "arg2" "arg3" "arg4"
This would facilitate a lot in scripting. Just a users suggestion.
2011/8/29 Ken Nielson <knielson at adaptivecomputing.com>:
> ----- Original Message -----
>> From: "Roy Dragseth" <roy.dragseth at cc.uit.no>
>> To: "Torque Users Mailing List" <torqueusers at supercluster.org>
>> Sent: Saturday, August 27, 2011 2:46:53 PM
>> Subject: [torqueusers] qsub patch for job script arguments.
>> I just discovered the -F option implemented in v3.0.2 and thought it
>> was a
>> very nice idea. Soon I discovered the awkwardness imposed by the
>> need for
>> quotes around the argument. In my experience it is very hard to
>> explain to
>> users the difference between single and double quotes and so on.
>> Maybe a
>> better way to solve this is to make everything between a "--" flag
>> and end of
>> line become job script arguments. This is the way getopt suggests to
>> do these
>> kind of things.
>> Attached is a patch that implements the idea above.
>> qsub run.sh -- arg1 arg2
>> will start
>> run.sh arg1 arg2
>> The equivalent would be
>> qsub -F"arg1 arg2" run.sh
>> but no need for quotes.
>> Anyway, just a suggestion.
>> The Computer Center, University of Tromsø, N-9037 TROMSØ Norway.
>> phone:+47 77 64 41 07, fax:+47 77 64 41 00
>> Roy Dragseth, Team Leader, High Performance Computing
>> Direct call: +47 77 64 62 56. email: roy.dragseth at uit.no
> Thanks for the patch.
> I think the -F option will cause some confusion for everyone at first because it is a different syntax than what people are used to. The way Maui and Moab interpret the command line is also a factor in this. Both programs would need to know the -F meant everything else following are arguments.
> Even so I am open to better practices. What does everyone else think.
> torqueusers mailing list
> torqueusers at supercluster.org
Daniel Fernando Coimbra
Grupo de Estrutura Eletrônica Molecular
Departamento de Química
Universidade Federal de Santa Catarina
More information about the torqueusers