Re: [HTCondor-users] Mpirun on htcondor 9

[John M Knoeller <johnkn@xxxxxxxxxxx>]

It looks like you are using 4 threads, so your submit file should be requesting at least 4 CPUs

You should add this to your submit file.

 

request_cpus = 4

 

As for the error, It sounds like you need to transfer more than just the P4.in file.

when you run this command interactively and it works for you, is there an rsh directory in the current working directory?  If so you could try adding that directory to your transfer input list

 

transfer_input_files = P4.in rsh

 

If that does not work, you might have better luck using the fork launcher rather than the rsh launcher.

 

-tj

 

From: HTCondor-users <htcondor-users-bounces@xxxxxxxxxxx> On Behalf Of Alejandro AcuÃa
Sent: Thursday, March 17, 2022 8:13 AM
To: htcondor-users <htcondor-users@xxxxxxxxxxx>
Subject: [HTCondor-users] Mpirun on htcondor 9

 

Hi all.

New blend for me. MPI and Condor.

I'm trying to submit first simple mpi test job under Ubuntu htcondor (9.6.0). The command is:

 

mpirun -np 4 pw.x -in P4.in > P4xx.out 

 

When I run it ah simple terminal submit, results appears ok (it takes 2 minutes). Data is correctly stored in the .out file.

But, if I put that command into sh file and submit via HTC under vanilla universe (I think it's possible), this error appears:

 

--------------------------------------------------------------------------

The value of the MCA parameter "plm_rsh_agent" was set to a path

that could not be found:

 

  plm_rsh_agent: ssh : rsh

 

Please either unset the parameter, or check that the path is correct

--------------------------------------------------------------------------

 

Sub file:

 

############################################################

## submit description file for

## MPI under the vanilla universe; CDE used

############################################################

universe = vanilla

output     = log/out.out

error       = log/error.err

log          = log/log.log

executable = mpi.sh

should_transfer_files = yes

when_to_transfer_output = on_exit

transfer_input_files = P4.in

 

queue

 

Help please! How can I configure the ssh so that the condor finds the correct path? 

 

Thanks

Ale