Hello John,
You need to specify both --enable-static and --disable-shared to do a static
build (not sure why, perhaps someone else can fill us in on that)...
The logs indicate the launch is failing trying to start orted on the backend
node... probably due to shared library dependencies.
You might try doing a bpsh <node> ldd orted
And check that the libraries resolve / and or rebuild with the indicated
configure option.
Regards,
Tim
John Ouellette wrote:
Hi,
I'm having problems with getting code (specifically ASC FLASH) to run on our
bproc-based cluster using Open-MPI.
Our cluster is based on bproc 4.0.0.pre8 and Linux 2.6.9 (vanilla) and we're
using Intel v9.0 compilers. OpenMPI compiles without errors on this setup.
I've attached a compressed file containing the output of config.log, my
machines file, and the output of ompi_info (the make.log file was too large
and made the tar file 120KB -- larger than the 100KB limit on the mailing
list).
The problem I'm having is detailed in the 'log3.out' file, also contained in
the attached tar file. Specifically, when I run FLASH with the following
command:
mpirun --hostfile machines.openmpi -np 4 ./flash2
I get a bzillion errors of the form:
ORTE_ERROR_LOG: Not found in file rmgr_urm_component.c at line 190
This eventually goes away and, just before the process dies, I also get many
errors like this:
demeter:27162] [0,0,0] ORTE_ERROR_LOG: Error in file pls_bproc.c at line 582
[demeter:27162] [0,0,0] ORTE_ERROR_LOG: Error in file pls_bproc.c at line 838
[demeter:27162] [0,0,0] ORTE_ERROR_LOG: Error in file rmgr_urm.c at line 177
and complaints about not being able to find the file help_pls_bproc.txt, which
actually does exist in the location where OpenMPI was installed.
Any insight would be greatly appreciated.
John Ouellette
------------------------------------------------------------------------
_______________________________________________
users mailing list
us...@open-mpi.org
http://www.open-mpi.org/mailman/listinfo.cgi/users
