Am 13.12.2012 um 18:16 schrieb Dave Love:

> Reuti <[email protected]> writes:
> 
>>> Two question here:
>>> 1.-) if set to false, the usage is not used but the measured?
>> 
>> During execution of the job you can see it in `qstat -j <job_id>`, but in 
>> `qacct -j <job_id>` the reserved one is recorded.
> 
> I think the qstat usage is treated the same way; I documented that after
> checking.  [I recently modified qacct to print the resource request
> too.]
> 
>>> 2.-) It talks about "reserved resources", but -l talks about
>>> "requested" resources. So, do I have to force users to reserver
>>> requested resources?
>> 
>> Two words for the same. While it was most likely a clear statement in
>> the past, with the introduction the advance reservation it could be
>> misleading. But changing the name of the settings to read
>> "*_REQUESTED_*" could make former cluster configurations invalid. You
>> could even argue it should be "*_GRANTED_*".
> 
> I'd have said they're not the same generally.  The reserved granted
> resources are a subset of the requested ones.  Even an h_vmem request,
> say, could be soft.  (This is relevant for resource reservation as well
> as advance reservation.)

True for soft requests of booleans/strings (where is it set e.g. to prefer a 
certain type of machine, but others would be accepted too), but the soft 
request of h_vmem, h_cpu and h_rt are different, as they are more treated like 
to generate a warning if passed. And consumables can't be soft.

-- Reuti


> -- 
> Community Grid Engine:  http://arc.liv.ac.uk/SGE/


_______________________________________________
users mailing list
[email protected]
https://gridengine.org/mailman/listinfo/users

Reply via email to