, it
might be confused by that.
I see that folks found their ways around the behaviour of Slurm … now
I'm only missing an explanation for why Slurm tries so hard to mess
with the job environment so that people feel the need to work around it …
Alrighty then,
Thomas
--
Dr. Thomas Orgis
Universität
$@ --pty -u ${SHELL} -i -l
, or do you intend to have quoted arguments being mangled? It might not
matter in practice for srun, as you rarely need quote something in its
arguments, but have it deeply ingrained to always use $@ for the
argument list.
Alrighty then,
Thomas
--
Dr. Thomas Orgis
for this hint.
Alrighty then,
Thomas
--
Dr. Thomas Orgis
Universität Hamburg
RRZ / Zentrale Dienste / HPC
Schlüterstr. 70
20146 Hamburg
Tel.: 040/42838 8826
Fax: 040/428 38 6270
smime.p7s
Description: S/MIME cryptographic signature
to
the wrong place) and add a suggestive message to the report.
I'd like to do the same with Slurm, but so far I didn't see a way to
write that kind of info to the job output file automatically.
Alrighty then,
Thomas
--
Dr. Thomas Orgis
Universität Hamburg
RRZ / Zentrale Dienste / HPC
ch works nicely.
Though one still would like to do something about that prompt loop when
there is no terminal to get input from. Maybe I can prepare a patch
when I am less busy with getting things running.
Alrighty then,
Thomas
--
Dr. Thomas Orgis
Universität Hamburg
RRZ / Zentrale Dienste
sages as fast as it
can here instead of coming to some sensible conclusion. Is that desired?
Alrighty then,
Thomas
--
Dr. Thomas Orgis
Universität Hamburg
RRZ / Zentrale Dienste / HPC
Schlüterstr. 70
20146 Hamburg
Tel.: 040/42838 8826
Fax: 040/428 38 6270
jobs, so that a job can use multiple
cores. Though, if you're I/O bound, I don't see the point.
Alrighty then,
Thomas
--
Dr. Thomas Orgis
Universität Hamburg
RRZ / Zentrale Dienste / HPC
Schlüterstr. 70
20146 Hamburg
Tel.: 040/42838 8826
Fax: 040/428 38 6270
smime.p7s
Description: S/MIME cryptographic signature
Am Thu, 24 Sep 2015 09:27:01 -0700
schrieb Thomas Orgis <thomas.or...@uni-hamburg.de>:
> is there a way to tell the Slurm daemons on the controlling server
> _not_ to listen to commands changing configuration from clients?
Given the traffic on this list on other topics, may I assum
f machines / root
accounts already using ssh, which also does not rely on nobody
attempting IP spoofing.
I'll see if the path via munge works. Thanks for the comments, both of
you.
Alrighty then,
Thomas
--
Dr. Thomas Orgis
Universität Hamburg
RRZ / Zentrale Dienste / HPC
Schlüterstr. 70
20146 Hambur
ocumentation on that (any configuration aspects of munge
or auth/munge). Or do you mean to patch the source? That's of course
also a possibility.
Alrighty then,
Thomas
--
Dr. Thomas Orgis
Universität Hamburg
RRZ / Zentrale Dienste / HPC
Schlüterstr. 70
20146 Hamburg
Tel.: 040/42838 8826
Fax: 040/428 38 6270
g the same route as
reconfiguration traffic. Am I overlooking something?
Alrighty then,
Thomas
--
Dr. Thomas Orgis
Universität Hamburg
RRZ / Zentrale Dienste / HPC
Schlüterstr. 70
20146 Hamburg
Tel.: 040/42838 8826
Fax: 040/428 38 6270
it is not unreasonable to have those
configuration and health checks separately, UNIX-style.
Alrighty then,
Thomas
--
Dr. Thomas Orgis
Universität Hamburg
RRZ / Zentrale Dienste / HPC
Schlüterstr. 70
20146 Hamburg
Tel.: 040/42838 8826
Fax: 040/428 38 6270
smime.p7s
Description: S/MIME cryptographic signature
ionality is present in released versions of
Slurm and folks might depend on it, I at least hope that configration
options to deactivate the intermediate to higher magic will be welcomed
to be included in the codebase. Not that I have patches ready, but I
might find time to prepare them.
Alrighty then,
cate them about
the option of setting TMPDIR to the correct one.
Alrighty then,
Thomas
--
Dr. Thomas Orgis
Universität Hamburg
RRZ / Zentrale Dienste / HPC
Schlüterstr. 70
20146 Hamburg
Tel.: 040/42838 8826
Fax: 040/428 38 6270
smime.p7s
Description: S/MIME cryptographic signature
sing issues in future).
Is that way of running MPI jobs in Slurm not supported?
Alrighty then,
Thomas
--
Dr. Thomas Orgis
Universität Hamburg
RRZ / Basisinfrastruktur / HPC
Schlüterstr. 70
20146 Hamburg
Tel.: 040/42838 8826
Fax: 040/428 38 6270
smime.p7s
Description: S/MIME cryptographic signature
ignored in our test cluster. If that does not yield, I'll follow
your example.
Alrighty then,
Thomas
--
Dr. Thomas Orgis
Universität Hamburg
RRZ / Basis-Infrastruktur / HPC
Schlüterstr. 70
20146 Hamburg
Tel.: 040/42838 8826
Fax: 040/428 38 6270
smime.p7s
Description: S/MIME cryptographic signature
is ready for it. I do want to avoid unnecessary
job aborts due to bad timing of the periodic checks and also
unnecessary checks as such while a job is still filling the node and
the check might even interfere.
Alrighty then,
Thomas
--
Dr. Thomas Orgis
Universität Hamburg
RRZ / Basis
Am Fri, 20 Jan 2017 19:58:20 +0100
schrieb "Dr. Thomas Orgis" <thomas.or...@uni-hamburg.de>:
> So I know my check works, but the failure in the prolog is without
> consequence. It is not reported anywhere, apparently.
I need to correct that. The prolog failure indeed is rep
or sub-optimal parallelization. Please consider optimizing
this type of job.
This used to be easy with Torque … just echo >&2 in the epilogue script
… is it still a challenge with Slurm?
Alrighty then,
Thomas
--
Dr. Thomas Orgis
Universität Hamburg
RRZ / Zentrale Dienste / HPC
Schl
have to think about how much interference with the user's scripts we
want to afford.
Alrighty then,
Thomas
--
Dr. Thomas Orgis
Universität Hamburg
RRZ / Zentrale Dienste / HPC
Schlüterstr. 70
20146 Hamburg
Tel.: 040/42838 8826
Fax: 040/428 38 6270
smime.p7s
Description: S/MIME cryptographic signature
p in the log files? What's the reasoning?
I might have to start hacking on this myself in the Slurm sources, but
would like to know if any outcome of this would be considered for
inclusion.
Alrighty then,
Thomas
--
Dr. Thomas Orgis
Universität Hamburg
RRZ / Zentrale Dienste / HPC
Schlüterstr. 70
2
ut OpenMPI, I figured I have
to downgrade Slurm when Intel MPI does not work properly anymore. Even
if this is ultimately Intel MPI's fault, this would be a strong reason
for us to keep Slurm at the older version for the whole lifetime of our
cluster in order to support the existing binaries.
Alr
you'd get when logging into the compute node is dangerous IMHO. Another
good point about such batch scripts with modules is that they document
what enviroment they are designed for, perhaps down to the exact
version of programs they need.
Alrighty then,
Thomas
--
Dr. Thomas Orgis
Universität
23 matches
Mail list logo