Is `/tmp/tester` available on nodes `nid00021` and `nid00020`? `srun` attempts to execute the command on the execution node.
Or is `tester` a batch script? In that case you should run it using sbatch, which reads the script from the submission nodes and distributes the script to the allocated nodes when run. HTH Michael On Wed, Sep 23, 2015 at 11:58 AM, Jim Robanske <[email protected]> wrote: > > I'm having a problem that's rather confusing to me: > > > /opt/slurm/default/bin/srun --jobid=65581 --exclusive -n 6 -w > nid[00020-00021] /tmp/tester > slurmstepd: execve(): /tmp/tester: No such file or directory > slurmstepd: execve(): /tmp/tester: No such file or directory > slurmstepd: execve(): /tmp/tester: No such file or directory > slurmstepd: execve(): /tmp/tester: No such file or directory > slurmstepd: execve(): /tmp/tester: No such file or directory > slurmstepd: execve(): /tmp/tester: No such file or directory > srun: error: nid00021: tasks 3-5: Exited with exit code 2 > srun: Terminating job step 65581.0 > srun: Job step aborted: Waiting up to 2 seconds for job step to finish. > srun: error: nid00020: tasks 0-2: Exited with exit code 2 > > > ls -l /tmp/tester > -rwxr-xr-x 1 root root 3802171 Sep 23 12:35 /tmp/tester > > > Any clues as to what is going wrong here??? > > Thanks in advance. >
