Hello,
I have OSCAR 5.0 on SLES10SP1 (x86_64) but still get 2 errors at the
last step if I check the cluster. The logs are below
The last and biggest change I made was the switch from OpenMPI 1.1.1 to
OpenMPI 1.2.5 now. After the RPM build I changed the configuration that
OSCAR can use the new package.
The build system is also the testing system. So it isn't a clean (fresh)
SLES10SP1 and OSCAR 5.0. I think it is always a problem to get package
debendency problem and other stuff if it isn't a clean system correct.
I have build OpenMPI from the official OpenMPI src rpm. It has tm
support.
# ompi_info | grep tm
MCA memory: ptmalloc2 (MCA v1.0, API v1.0, Component
v1.2.5)
MCA ras: tm (MCA v1.0, API v1.3, Component v1.2.5)
MCA pls: tm (MCA v1.0, API v1.3, Component v1.2.5)
About the LAM problem I read
(http://www.mail-archive.com/[email protected]/msg04656.html)
that it could be a problem if the package was build on a host without torque.
But this problem is very old and from OSCAR 4.2. I tried it with torque but
didn't help me to solve it.
There were no errors at the lam and openmpi RPM build.
And in every log I had a look I can not find more information what can
cause this 2 problems.
The PVM and MPICH tests are passed so it couldn't be a possible switcher
problem I think.
For every help I would be glad. If you need more information or a
logfile just let me know.
Performing root tests...
Maui service check:maui [PASSED]
TORQUE node check [PASSED]
TORQUE service check:pbs_server [PASSED]
/home mounts [PASSED]
Preparing user tests...
Performing user tests...
SSH ping test [PASSED]
SSH server->node [PASSED]
SSH node->server [PASSED]
LAM/MPI (via TORQUE) [FAILED]
PVM (via TORQUE) [PASSED]
MPICH (via TORQUE) [PASSED]
Ganglia setup test [PASSED]
Ganglia node count test [PASSED]
TORQUE default queue definition [PASSED]
TORQUE Shell Test [PASSED]
Open MPI (via TORQUE) [FAILED]
qdel: Request invalid for state of job 106.sles10oscar
Run APItests...
Running Installation tests for pvm
[PASS] 2008-01-23 09:54:49 pvmd-path-ls.apt
[PASS] 2008-01-23 09:54:49 envvar-pvm_arch.apt
[PASS] 2008-01-23 09:54:49 envvar-pvm_root.apt
[PASS] 2008-01-23 09:54:49 envvar.apb
[PASS] 2008-01-23 09:54:49 pvmd-path-which.apt
[PASS] 2008-01-23 09:54:49 modulecmd-path-ls.apt
[PASS] 2008-01-23 09:54:49 pvm-module-list.apt
[PASS] 2008-01-23 09:54:49 pvm-module-show-pvm_rsh.apt
[PASS] 2008-01-23 09:54:49 pvm-module-show-pvm_arch.apt
[PASS] 2008-01-23 09:54:49 pvm-module-show-pvm_root.apt
[PASS] 2008-01-23 09:54:49 pvm-module-show.apb
[PASS] 2008-01-23 09:54:49 pvm-module.apb
[PASS] 2008-01-23 09:54:49 install_tests.apb
There are 2 failed/skipped tests (see above).
Please check for .err and .out files in /home/oscartst/<package>.
...Hit <ENTER> to close this window...
--------------------------------------------------------------------------------
sles10oscar:/home/oscartst/lam # cat lamtest.out
Running LAM/MPI test
sles10oscar:/home/oscartst/lam # cat lamtest.err
ERROR: LAM/MPI does not appear to have the tm boot SSI module!
This test script will now abort.
sles10oscar:/home/oscartst/lam #
--------------------------------------------------------------------------------
sles10oscar:/home/oscartst/openmpi # cat openmpitest.err
[oscarnode2:08400] pls:tm: failed to poll for a spawned proc, return
status = 17002
[oscarnode2:08400] [0,0,0] ORTE_ERROR_LOG: In errno in file rmgr_urm.c
at line 462
[oscarnode2:08400] mpiexec: spawn failed with errno=-11
sles10oscar:/home/oscartst/openmpi # cat openmpitest.out
Running Open MPI test
Open MPI appears to have TM suppport. Yippee!
--> MPI C bindings test:
TEST FAILED!
Commands: cp cpi.c /tmp/openmpi-test && cd /tmp/openmpi-test && mpicc
cpi.c -o openmpi-cpi && cp openmpi-cpi /home/oscartst/openmpi &&
cd /home/oscartst/openmpi && mpiexec
-machinefile /var/spool/pbs/aux//106.sles10oscar -n 2 openmpi-cpi
sles10oscar:/home/oscartst/openmpi #
--
Mit freundlichen Grüßen / Best regards
Dominik Schips
Tel.: +49 (0)21 61 - 46 43-112
Fax: +49 (0)21 61 - 46 43-100
credativ GmbH, HRB Mönchengladbach 12080
Hohenzollernstr. 133, 41061 Mönchengladbach
Geschäftsführung: Dr. Michael Meskes, Jörg Folz
-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Oscar-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/oscar-devel
