Hello Collin,

If you can do it, could you try to ssh into one of the nodes where a hello_c 
process is running and attach to it with a debugger and get a traceback?

Howard

From: 'Collin Strassburger' via Open MPI users <[email protected]>
Reply-To: "[email protected]" <[email protected]>
Date: Tuesday, December 9, 2025 at 1:19 PM
To: Open MPI Users <[email protected]>
Subject: [EXTERNAL] [OMPI users] Multi-host troubleshooting

Hello,

I am dealing with an odd mpi issue that I am unsure how to continue diagnosing.

Following the outline given by: 
https://www.open-mpi.org/faq/?category=running#diagnose-multi-host-problems<https://urldefense.com/v3/__https:/www.open-mpi.org/faq/?category=running*diagnose-multi-host-problems__;Iw!!Bt8fGhp8LhKGRg!DRYP94rDXIEmdklDrsyM2wrE-tkiXvRFPv3wNgs_QqJDo2u6ltIg7OwPot_gKi3FSnGX1TdHa9QBl6J5JTaV$>,
 steps 1-3 complete without any issues
i.e. ssh remotehost hostname works
Paths include the nvidia hpcx paths when checked both with ssh and mpirun
Mpirun --host node1,node2 hostname works correctly
Mpirun --host node1,node2 env | grep –i path yields identical paths which 
include the paths required by hpcx
(This is all through passwordless login)

Step 4 calls to run mpirun --hosts node1,node2 hello_c.  I have locally 
compiled the code and confirmed that it works on each machine individually.  
The same code is shared between the machines.  However, it does not run across 
multiple hosts at once.  It simply hangs until Ctrl-C’d.  I have attached the 
--mca plm_base_verbose 10 logs; while I don’t see anything in them, I am not 
well versed enough in OpenMPI to think that I understand the full implications 
of it all.

Notes:
No firewall is present between the machines (minimal install is the base, so 
ufw and iptables are not present by default and have not yet been installed)
Journalctl does not report any errors.
The machines have identical hardware and utilized the same configuration script.
Calling “mpirun --hosts node1,node2 mpirun --version” returns identical results
Calling “mpirun --hosts node1,node2 env | grep -i path” returns identical 
results

OS: Ubuntu 24.04 LTS
OMPI: 4.1.7rc1 from Nvidia HPCX
Configure options:
    --prefix=${HPCX_HOME}/ompi \
    --with-hcoll=${HPCX_HOME}/hcoll \
    --with-ucx=${HPCX_HOME}/ucx \
    --with-platform=contrib/platform/mellanox/optimized \
    --with-tm=/opt/pbs/ \
    --with-slurm=no \
    --with-pmix \
    --with-hwloc=internal

I’m rather at a loss on what to try/check next.  Any thoughts on how to 
continue troubleshooting this issue?

Warm regards,
Collin Strassburger (he/him)

________________________________
The information contained in this e-mail and any attachments from Bihrle 
Applied Research may contain confidential and/or proprietary information, and 
is intended only for the named recipient to whom it was originally addressed. 
If you are not the intended recipient, any disclosure, distribution, or copying 
of this e-mail or its attachments is strictly prohibited. If you have received 
this e-mail in error, please notify the sender immediately by return e-mail and 
permanently delete the e-mail and any attachments.
To unsubscribe from this group and stop receiving emails from it, send an email 
to 
[email protected]<mailto:[email protected]>.

To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].

Reply via email to