[torqueusers] Re: Evaluating DRMAA + Torque

Darren Platt darren at 23andme.com
Tue Jun 3 15:02:46 MDT 2008


Thanks!  One of the attractions is the portability, so we don't have to
recode if we outgrow or change schedulers.  Part of the problem seems to be
the abstraction.  For example, setting the outputPath  field in the job to
myhost:/some/path/to/some/file.txt    didn't deliver the result to the
headnode in the way -o   would.  The problem seemed to be in how the drmaa
layer was rewriting the path,  assumptions about whether the intermediate
path /some/path / etc existed on the node vs the submit host.    So I assume
most people are using either the shell or torque api to submit then?

Darren

On Tue, Jun 3, 2008 at 1:14 PM, Joshua Bernstein <
jbernstein at penguincomputing.com> wrote:

> Hi Darren,
>
>        While I haven't used DRMAA with TORQUE directly, my understanding of
> DRMAA is that it is still in its infancy. Though my other limited experience
> with DRMAA has shown that it provides a nice abstracting between your
> software and a scheduler, a layer of portability if you well.
>        Though that said, different DRMAA layers are at different states of
> maturity, so some have different gotchas here there.
>
> -Joshua Bernstein
> Software Engineer
> Penguin Computing
>
>
>
> Darren Platt wrote:
>
>> Hi,
>>
>> we've been testing out Torque + DRMAA combination and using DRMAA-py that
>> was written for SGE.  I've
>> encountered a few issues and was wondering if there are many users out
>> there using Torque+DRMAA.  The underlying
>> Torque implementation seems to be behaving correctly, but I've had a lot
>> of issues trying to get input/output files to stage
>> correctly using scp.  I ran into some slight problems compiling the python
>> wrappers as well that required a small addition to
>> the torque/DRMAA headers.  Has anyone used DRMAA-py with Torque?   Any
>> comments appreciated.   We're testing
>> a variety of different schedulers and I was impressed with the relative
>> ease of installation for Torque.  The docs were a little out
>> of date in places, but overall good.  We're testing 2.3.0,
>>
>> Thanks,
>>
>> --
>> Darren Platt
>> Senior Director, Research
>> 23andMe, inc
>>
>>
>> ------------------------------------------------------------------------
>>
>> _______________________________________________
>> torqueusers mailing list
>> torqueusers at supercluster.org
>> http://www.supercluster.org/mailman/listinfo/torqueusers
>>
>


-- 
Darren Platt
Senior Director, Research
23andMe, inc
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.supercluster.org/pipermail/torqueusers/attachments/20080603/0bc317f4/attachment.html


More information about the torqueusers mailing list