Gstat is showing Total Grid jobs as 3703 at Oxford. I'm not sure how it's getting to this total, we have a small ce(02) which drives it's own torque server with 16 job slots, and t2ce04 and 06 which front the same batch server with 1344 jobs slots (although quite a few are offline today).
The qstats below show a total running jobs of 15 + 1012 = 1027
With 167 + 1055 = 1222 jobs queuing so total jobs = 2249 (If that's what total jobs is really trying to show)
Cheers Pete
[root@t2ce02 ~]# qstat -q
server: t2ce02.physics.ox.ac.uk
Queue Memory CPU Time Walltime Node Run Que Lm State
---------------- ------ -------- -------- ---- --- --- -- -----
mediumfive -- 24:00:00 36:00:00 -- 2 77 -- E R
expressfive -- 01:00:00 01:30:00 -- 0 0 -- E R
longfive -- 48:00:00 72:00:00 -- 2 82 -- E R
shortfive -- 12:00:00 18:00:00 -- 11 8 -- E R
----- -----
15 167
[root@t2ce04 ~]# qstat -q
server: t2torque02.physics.ox.ac.uk
Queue Memory CPU Time Walltime Node Run Que Lm State
---------------- ------ -------- -------- ---- --- --- -- -----
shortfive -- 12:00:00 18:00:00 -- 10 1 -- E R
expressfive -- 01:00:00 01:30:00 -- 0 0 -- E R
mediumfive -- 24:00:00 36:00:00 -- 39 60 -- E R
longfive -- 48:00:00 72:00:00 -- 963 994 -- E R
----- -----
1012 1055
[root@t2ce06 ~]# qstat -q
server: t2torque02.physics.ox.ac.uk
Queue Memory CPU Time Walltime Node Run Que Lm State
---------------- ------ -------- -------- ---- --- --- -- -----
shortfive -- 12:00:00 18:00:00 -- 10 1 -- E R
expressfive -- 01:00:00 01:30:00 -- 0 0 -- E R
mediumfive -- 24:00:00 36:00:00 -- 39 60 -- E R
longfive -- 48:00:00 72:00:00 -- 963 994 -- E R
----- -----
1012 1055
--
----------------------------------------------------------------------
Peter Gronbech Senior Systems Manager and Tel No. : 01865 273389
GridPP Project Manager Fax No. : 01865 273418
Department of Particle Physics,
University of Oxford,
Keble Road, Oxford OX1 3RH, UK E-mail : [log in to unmask]
----------------------------------------------------------------------
-----Original Message-----
From: Testbed Support for GridPP member institutes [mailto:[log in to unmask]] On Behalf Of Stephen Burke
Sent: 07 May 2012 12:51
To: [log in to unmask]
Subject: Re: gstat job counting
Testbed Support for GridPP member institutes [mailto:TB-
> [log in to unmask]] On Behalf Of Christopher J. Walker said:
> However, it means that the grid does the maths[1]. I can say I have A
> nodes with B hepspec and C nodes with D hepspec. When I add some
> additional nodes, I can easily change it without digging out the
> spreadsheet.
Surely that only works if you happen to have the same number of CEs as you have types of node?
> To answer Stephen's point, jobs can't rely on which node they end up
> on
> - so job requirements might not be met. However, the nodes we have
> have fairly similar hepspec scores, and RAM/jobslot.
In that case there isn't much advantage to publishing them separately ...
Aside from this discussion, it would still be interesting to know the answer to the original question, i.e. if sites with multiple identical CEs still see overcounting of jobs in gstat.
Stephen
|