I would like to certainly express my interest too.
Our plans are to move to SLURM by 2013 Q2 (work to be started by Q1) and
having a place to bi-directionally share experiences would be great.
Thanks,
BR/Pablo
On 11/14/2012 03:53 PM, Mario Kadastik wrote:
>>> I was curious as to the progress on SLURM integration? If I remember correctly it was planned for EMI-3 release, but are there beta releases or plans for first adaptors as we're really really pissed off about our current Torque+maui workings.
>> Status of SLURM support by CREAM was reported at the EGI Technical Forum (https://indico.egi.eu/indico/getFile.py/access?contribId=37&sessionId=46&resId=0&materialId=slides&confId=1019)
>>
>> Quoting from the slides:
>> "CREAM status:
>> BLAH script => OK
>> Infoprovider => Work-in-progress
>> APEL Sensors => Work-in-progress"
>>
>> The testing results reported in the slides are very promising.
>>
>> One PL-Grid (ICM) site is running CREAM on SLURM in production, and they already expressed interest in being an early adopter.
>>
>> The CREAM PT is collecting expressions of interests now from sites for testing, feedback, and development tasks (PT contact: [log in to unmask])
>>
>> EGI can support a community forum as a mechanism to start exchanging expertise between site administrators, and can provide its wiki for exchange of information/documentation. Are you interested in this? setting up a discussion forum on SLURM is straighforward.
>
> Yes I would be very interested in this. We can always set up SLURM as a second scheduler with a dedicated CREAM initially for testing and then later can move it to production if it works etc. As torque and maui are configured to schedule jobs based on workernode load we could even run SLURM and PBS at the same time on same workernodes for a while. So definitely interested.
>
> Mario Kadastik, PhD
> Researcher
>
> ---
> "Physics is like sex, sure it may have practical reasons, but that's not why we do it"
> -- Richard P. Feynman
|