Also check the script has exec permission El 16/05/2012 06:38, "[email protected]" <[email protected]> escribió:
> check that the user/uid is valid on the allocated compute node > -- > Sent from my Android phone. Please excuse my brevity and typos. > > andre roy <[email protected]> wrote: >> >> Hey folks, >> >> I'm just trying to setup slurm using the new Ubuntu LTS distribution and >> am running into a strange issue I haven't seen before. I'm trying to run a >> simple script whose contents are: >> >> --------------- >> #!/bin/bash >> >> hostname > results.dat >> >> ----------------- >> >> this file runs succesfully with srun, however running it with sbatch >> fails. I've narrowed the problem to the cluster node. The >> /var/log/slurm-llnl/slurmd.log reads: >> >> ------------- >> Launching batch job 30 for UID 10000 >> execve(): /var/lib/slurm-llnl/slurmd/job00030/slurm_script: Permission >> Denied >> sending REQUEST_COMPLETE_BATCH_SCRIPT, error:0 >> done with job >> ------------------- >> >> I've set the folder owner to slurm:slurm for /var/lib/slurm-llnl and all >> subfolders, I'm out of ideas as to why permission was denied, any ideas? >> >> Thanks, I hope someone has a solution :) >> >> Cheers, >> >> André > >
