[PBS-USERS] mpiexec with OpenPBS_2.3.16

[Anupam Sharma]

Hi Pete,

	I am still having trouble.  Here is the error I get:


This job has allocated 2 nodes
p0_29743: (0.007079) send_message: to=1; invalid conn type=5
p0_29743:  p4_error: subtree_broadcast_p4 failed, type=: 1010101010

mpiexec: Warning: main: task 0 exited with status 1 (raw 0x1).


thanks,
anupam

----------------- your mail ----------------

| anupam-sharma at psu.edu said:
| > 	I am having trouble making mpiexec work with OpenPBS on my mini
| > cluster of 3 dual processor nodes.  I am able to compile mpiexec without
| > any problems and also able to submit the job using a simple script which
| > looks like this.
| >
| > #!/bin/sh
| > #PBS -N TEST2
| > #PBS -e MYJOB.err
| > #PBS -o MYJOB.log
| > #PBS -l nodes=1:ppn=1
| >
| > # This job's working directory
| > echo Working directory is $PBS_O_WORKDIR
| > cd $PBS_O_WORKDIR
| >
| > echo Running on host `hostname`
| > echo Time is `date`
| > echo Directory is `pwd`
| > echo This jobs runs on the following processors:
| > echo `cat $PBS_NODEFILE`
| > # Define number of processors
| > NPROCS=`wc -l < $PBS_NODEFILE`
| > echo This job has allocated $NPROCS nodes
| >
| > /usr/local/bin/mpiexec cpi
| >
| > # ----------------------  END -----------------------
| >
| > The problem I am having is that the job just hangs in the middle of the
| > run and never completes!!
| > I have installed mpiexec on /usr/local/bin which is mounted across the
| > nodes.
| >
| > I looked up on the web and there were some patches for PBS to support
| > mpiexec however none say anything about this type of problem.
|
| Were you able to resolve this?  You can add "-v" (one or many) arguments
| to mpiexec to see exactly what it does.  Also make sure you are using
| the right communication library.  Mpiexec must be told if your code is,
| say, mpich/gm or mpich/p4.
|
| Followup to mpiexec at osc.edu, not pbs-users, please.
|
| 		-- Pete
|