On Thursday, 30 March 2017, at 09:19:59 (-0700),
Chad Cropper wrote:
> Yes, I have seen these. Thanks. But the issue is I only want this to
> run when the node is empty. Our workload is very serial and most job
> sonly use 1,2,4 cores. So on a large node of 32 cores, we would have
> many jobs
A bit late, but thank you for the answer.
It doesn't help me since my problem is truly specific, when the queue is
full, the dependecy act weird and just try to send the job infinite times.
Anyway, I solved it using some loops in my scripts, not the best way,
but...
*Álvaro Ponce Cabrera.*
When the srun was ran the following output was given:
srun -n8 -l hostname
srun: error: Task launch for 1954.0 failed on node rhinonode01: User
not found on host
srun: error: Application launch failed: User not found on host
srun: Job step aborted: Waiting up to 32 seconds for job step to
Hi All
I have a QoS where the limit is set like this:
# sacctmgr show qos qos_bloggs
--trim
Name GrpTRESMins
qos_bloggs cpu=450
only the bloggs account is able to use this QoS, please trust me on this
I've checked 3 or 4 times
AccountQOS
AFAIK, sbatch and friends don't allow "preferences" when submitting jobs
to multiple "clusters/partitions". Theoretically you could define a
different QOS for each workload, but either that QOS would be valid for
all nodes/partitions, or it would be the same as submitting to a single
partition
You can define nodes with "features", and then at job submission time
require specific features.
E.g., if you had some nodes with ethernet only, some with QDR infiniband,
and some with FDR ib, you could define the QDR and FDR nodes to have
the feature qdr or fdr, respectively. Then, e.g. a
Hello again! I think I found a solution to the problem I reported in
this thread.
I have checked Slurm's source code and realized that in file
/slurm-17.02.1-2/src/plugins/accounting_storage/mysql/accounting_storage_mysql.c
the function used to create _assoc_table (create_cluster_tables)