Mmh,
I have only 6.2u5 around. I can only state, that for me it seems
working as it should. Maybe someone having your version of SGE can
perform the test too, to check whether there is any additonal setup
misconfigured, or whether it's a bug in this version.
Am 31.05.2011 um 21:24 schrie
Am 31.05.2011 um 21:29 schrieb Dave Love:
Reuti writes:
You can use a JSV to request a consumable complex when you notice the
"-sync y" option (or users are fair and do it on their own already).
Actually, you can't as -sync isn't available to jsv.
Thx, I wasn't aware of it. I can see a re
Reuti writes:
> Hi,
>
> Am 31.05.2011 um 18:14 schrieb Patterson, Ron (NIH/NLM/NCBI) [C]:
>
>> I've never played with JSV's before so I'll have to read up on them.
>> I seem to remember that they can cause performance issues if you are
>> not careful - is that right?
>
> well, it's suggested
Reuti writes:
> You can use a JSV to request a consumable complex when you notice the
> "-sync y" option (or users are fair and do it on their own already).
Actually, you can't as -sync isn't available to jsv.
qsub and jsv_script_interface at http://arc.liv.ac.uk/SGE/htmlman/
should now be
Reuti,
Thank you for your response and the helpful link. Unfortunately, I
cannot seem to make the scheduler work as described. Here is my
scheduler configuration:
[gladden@stuart ~]$ qconf -ssconf
algorithm default
schedule_interval 0:0:06
maxujobs
Hi,
Am 25.05.2011 um 17:30 schrieb Dave Love:
It seems that core binding currently isn't typically useful with MPI
jobs. The pe method needs exclusive node access and a fixed
allocation
rule that you need to know in advance, but then you could have the mpi
do it anyhow (e.g. openmpi paffini
There is a "Performance Considerations" section of the SGE docs which describes
writing efficient rules for resource quota sets - that's all I was referring to.
Our SGE cluster runs very large numbers of short jobs per day (many are too
short), so qmaster load is an important factor for me. I'l
Hi,
Am 31.05.2011 um 18:14 schrieb Patterson, Ron (NIH/NLM/NCBI) [C]:
I've never played with JSV's before so I'll have to read up on them.
I seem to remember that they can cause performance issues if you are
not careful - is that right?
well, it's suggested to use a compiled language like
Thanks Reuti.
I've never played with JSV's before so I'll have to read up on them. I seem to
remember that they can cause performance issues if you are not careful - is
that right? I know the same is true for RQS's, which we depend on quite a bit.
Ron
-Original Message-
From: Reuti [m
Am 31.05.2011 um 17:22 schrieb Patterson, Ron (NIH/NLM/NCBI) [C]:
Is there any way to limit the number of jobs using "-sync y" allowed
per user? I know I can limit the total number of client connections
allowed, and I can limit the total number of jobs by a group of
users. I want to cap the
I guess I should add, we are using SGE 6.2u5, from the last courtesy binaries
released by Sun/Oracle last year.
Ron
-Original Message-
From: Patterson, Ron (NIH/NLM/NCBI) [C]
Sent: Tuesday, May 31, 2011 11:22 AM
To: users@gridengine.org
Subject: Any way to limit "qsub -sync y" jobs per
Is there any way to limit the number of jobs using "-sync y" allowed per user?
I know I can limit the total number of client connections allowed, and I can
limit the total number of jobs by a group of users. I want to cap the number of
synchronous jobs per user, and total by cluster because too
12 matches
Mail list logo
