[Ekhi-users] CFM clusters partially down by Sunday August the 11th

[Inigo Aldazabal Mensa]

Hi all,

Due to the high temperatures expected for next Sunday, and in order to
be able to power off the computing nodes without affecting the service,
we are draining most of the CFM computer clusters' nodes so that jobs
expected to end (as by indicated in the --time option at the moment of
being run) after Sunday the 11th at 5am will not be run and will be kept
queued.

That is, you can still submit jobs as normal and, if they are "in
time", they'll be run. If not, they'll be kept in PENDING state and run
once we put everything up again.

Now, with this you see why you should  try to specify the job duration
--time in the slurm scripts and don't let slurm use the default value
one which usually is one week or so (it depends on every cluster
policies).

Remember that for any help or information regarding the clusters you can
write to the computing service common email 

CFM  Scientific Computing Service <hpc.cfm at ehu.eus>

and Irene and/or I will take care of your request.

Bests,

Iñigo

-- 
Iñigo Aldazabal Mensa, Ph.D.
HPC Computing Centre Manager / Scientific Computing Specialist
Centro de Física de Materiales (CSIC-UPV/EHU)
Paseo Manuel de Lardizabal, 5
20018 San Sebastian - Guipuzcoa
SPAIN

phone: +34-943-01-8780
e-mail: inigo.aldazabal at csic.es inigo.aldazabalm at ehu.eus
pgp key id: 0xDBCC8369