Discussion:
--task-prolog not called
Alexander Frolov
2014-07-15 12:22:30 UTC
Permalink
Hi!

I am using slurm 14.11.0-pre1 to strart job as following:

srun -N 1
--task-prolog=/home/frolo/grappa.master/build/Make+Release/prolog.sh --
bash /home/frolo/grappa.master/build/Make+Release/hello.sh

But prolog.sh is not called. What can be the reason?

Thank you!

Best,
Alex
-----------------------------------------------------------------------------------
prolog.sh:

echo "Hello (prolog)!"
echo "export TEST=1"
echo "print TTTT"

-----------------------------------------------------------------------------------
hello.sh

echo `hostname` "TEST=$TEST"
-----------------------------------------------------------------------------------
output:

A11 TEST=
-----------------------------------------------------------------------------------
Alexander Frolov
2014-07-15 13:40:37 UTC
Permalink
------------------
Ouput of slurmd:

slurmd: debug3: state for jobid 323: ctime:1405428718 revoked:0 expires:0
slurmd: debug: Checking credential with 128 bytes of sig data
slurmd: debug: task_p_slurmd_launch_request: 324.0 0
slurmd: debug: Calling /home/frolo/local/slurm/sbin/slurmstepd spank prolog
slurmstepd: test!!!
spank-prolog: Reading slurm.conf file:
/home/frolo/local/slurm/etc/slurm.conf
spank-prolog: Running spank/prolog for jobid [324] uid [1004]
spank-prolog: spank: opening plugin stack
/home/frolo/local/slurm/etc/plugstack.conf
slurmd: debug3: _rpc_launch_tasks: call to _forkexec_slurmstepd
slurmd: debug3: slurmstepd rank 0 (A11), parent rank -1 (NONE), children 0,
depth 0, max_depth 0
slurmd: debug3: _send_slurmstepd_init: call to getpwuid_r
slurmd: debug3: _send_slurmstepd_init: return from getpwuid_r
slurmd: debug3: _rpc_launch_tasks: return from _forkexec_slurmstepd
slurmd: debug: task_p_slurmd_reserve_resources: 324 0
slurmd: debug3: in the service_connection
slurmd: debug2: got this type of message 6004
slurmd: debug2: Processing RPC: REQUEST_SIGNAL_TASKS
slurmd: debug: Sending signal 995 to step 324.0
slurmd: debug3: in the service_connection
slurmd: debug2: got this type of message 6011
slurmd: debug2: Processing RPC: REQUEST_TERMINATE_JOB
slurmd: debug: _rpc_terminate_job, uid = 1004
slurmd: debug: task_p_slurmd_release_resources: 324
slurmd: debug: credential for job 324 revoked
slurmd: debug2: No steps in jobid 324 to send signal 18
slurmd: debug2: No steps in jobid 324 to send signal 15
slurmd: debug4: sent ALREADY_COMPLETE
slurmd: debug2: set revoke expiration for jobid 324 to 1405432351 UTS
slurmd: debug3: in the service_connection
slurmd: debug2: got this type of message 1008
------------------

/home/frolo/local/slurm/sbin/slurmstepd spank prolog -- is it correct?
According to sources it wont run task_prolog...
Post by Alexander Frolov
Hi!
srun -N 1
--task-prolog=/home/frolo/grappa.master/build/Make+Release/prolog.sh --
bash /home/frolo/grappa.master/build/Make+Release/hello.sh
But prolog.sh is not called. What can be the reason?
Thank you!
Best,
Alex
-----------------------------------------------------------------------------------
echo "Hello (prolog)!"
echo "export TEST=1"
echo "print TTTT"
-----------------------------------------------------------------------------------
hello.sh
echo `hostname` "TEST=$TEST"
-----------------------------------------------------------------------------------
A11 TEST=
-----------------------------------------------------------------------------------
Alexander Frolov
2014-07-15 15:04:40 UTC
Permalink
I dismiss my question, because just have found the issue:

#!/bin/bash -- must be added to prolog!

Best regs,
Alex
Post by Alexander Frolov
------------------
slurmd: debug3: state for jobid 323: ctime:1405428718 revoked:0 expires:0
slurmd: debug: Checking credential with 128 bytes of sig data
slurmd: debug: task_p_slurmd_launch_request: 324.0 0
slurmd: debug: Calling /home/frolo/local/slurm/sbin/slurmstepd spank prolog
slurmstepd: test!!!
/home/frolo/local/slurm/etc/slurm.conf
spank-prolog: Running spank/prolog for jobid [324] uid [1004]
spank-prolog: spank: opening plugin stack
/home/frolo/local/slurm/etc/plugstack.conf
slurmd: debug3: _rpc_launch_tasks: call to _forkexec_slurmstepd
slurmd: debug3: slurmstepd rank 0 (A11), parent rank -1 (NONE), children
0, depth 0, max_depth 0
slurmd: debug3: _send_slurmstepd_init: call to getpwuid_r
slurmd: debug3: _send_slurmstepd_init: return from getpwuid_r
slurmd: debug3: _rpc_launch_tasks: return from _forkexec_slurmstepd
slurmd: debug: task_p_slurmd_reserve_resources: 324 0
slurmd: debug3: in the service_connection
slurmd: debug2: got this type of message 6004
slurmd: debug2: Processing RPC: REQUEST_SIGNAL_TASKS
slurmd: debug: Sending signal 995 to step 324.0
slurmd: debug3: in the service_connection
slurmd: debug2: got this type of message 6011
slurmd: debug2: Processing RPC: REQUEST_TERMINATE_JOB
slurmd: debug: _rpc_terminate_job, uid = 1004
slurmd: debug: task_p_slurmd_release_resources: 324
slurmd: debug: credential for job 324 revoked
slurmd: debug2: No steps in jobid 324 to send signal 18
slurmd: debug2: No steps in jobid 324 to send signal 15
slurmd: debug4: sent ALREADY_COMPLETE
slurmd: debug2: set revoke expiration for jobid 324 to 1405432351 UTS
slurmd: debug3: in the service_connection
slurmd: debug2: got this type of message 1008
------------------
/home/frolo/local/slurm/sbin/slurmstepd spank prolog -- is it correct?
According to sources it wont run task_prolog...
On Tue, Jul 15, 2014 at 4:22 PM, Alexander Frolov <
Post by Alexander Frolov
Hi!
srun -N 1
--task-prolog=/home/frolo/grappa.master/build/Make+Release/prolog.sh --
bash /home/frolo/grappa.master/build/Make+Release/hello.sh
But prolog.sh is not called. What can be the reason?
Thank you!
Best,
Alex
-----------------------------------------------------------------------------------
echo "Hello (prolog)!"
echo "export TEST=1"
echo "print TTTT"
-----------------------------------------------------------------------------------
hello.sh
echo `hostname` "TEST=$TEST"
-----------------------------------------------------------------------------------
A11 TEST=
-----------------------------------------------------------------------------------
Continue reading on narkive:
Loading...