[mpich-discuss] mpicexec segmentation fault
Pavan Balaji
balaji at mcs.anl.gov
Thu Mar 31 20:05:22 CDT 2011
Hi Michael,
Are you setting any environment variables? We fixed a bug recently with
the HYDRA_BOOTSTRAP environment was being set. Can you try the 1.4rc1
version that we just released to see if the problem still exists?
Thanks,
-- Pavan
On 03/31/2011 05:34 PM, MICHAEL S DAVIS wrote:
> I recently downloaded mpich2-1.3.2p1 and compiled and installed the
> software with no errors.
>
> When I test run the test command
>
> [root at host bin]# ./mpiexec -np 1 -machinefile machines /bin/hostname
> Segmentation fault
> [root at host bin]#
>
> Here is the build information
>
> [root at host bin]# ./mpiexec -info
> HYDRA build details:
> Version: 1.3.2p1
> Release Date: Mon Feb 14 19:07:22 CST 2011
> CC: gcc
> CXX: c++
> F77: f77
> F90:
> Configure options: '--prefix=/opt/mpich2'
> '--disable-option-checking' 'CC=gcc' 'CFLAGS= -O2' 'LDFLAGS= '
> 'LIBS=-lrt -lpthread ' 'CPPFLAGS=
> -I/var/tmp/mpich2-1.3.2p1/src/mpl/include
> -I/var/tmp/mpich2-1.3.2p1/src/mpl/include
> -I/var/tmp/mpich2-1.3.2p1/src/openpa/src
> -I/var/tmp/mpich2-1.3.2p1/src/openpa/src
> -I/var/tmp/mpich2-1.3.2p1/src/mpid/ch3/include
> -I/var/tmp/mpich2-1.3.2p1/src/mpid/ch3/include
> -I/var/tmp/mpich2-1.3.2p1/src/mpid/common/datatype
> -I/var/tmp/mpich2-1.3.2p1/src/mpid/common/datatype
> -I/var/tmp/mpich2-1.3.2p1/src/mpid/common/locks
> -I/var/tmp/mpich2-1.3.2p1/src/mpid/common/locks
> -I/var/tmp/mpich2-1.3.2p1/src/mpid/ch3/channels/nemesis/include
> -I/var/tmp/mpich2-1.3.2p1/src/mpid/ch3/channels/nemesis/include
> -I/var/tmp/mpich2-1.3.2p1/src/mpid/ch3/channels/nemesis/nemesis/include
> -I/var/tmp/mpich2-1.3.2p1/src/mpid/ch3/channels/nemesis/nemesis/include
> -I/var/tmp/mpich2-1.3.2p1/src/mpid/ch3/channels/nemesis/nemesis/utils/monitor
> -I/var/tmp/mpich2-1.3.2p1/src/mpid/ch3/channels/nemesis/nemesis/utils/monitor
> -I/var/tmp/mpich2-1.3.2p1/src/util/wrappers
> -I/var/tmp/mpich2-1.3.2p1/src/util/wrappers'
> Process Manager: pmi
> Launchers available: ssh rsh fork slurm ll lsf
> sge persist
> Binding libraries available: hwloc plpa
> Resource management kernels available: none slurm ll lsf sge pbs
> Checkpointing libraries available:
> Demux engines available: poll select
> [root at host bin]#
>
> I also tried using the intel compilers and get the same result. The
> software builds and installs, but I can not run mpiexec without getting
> a segmentation fault. I have mpich1 running on this cluster and have no
> problems with it.
>
> I can ssh between nodes with no problems.
>
> I am running Red Hat Enterprise Linux WS release 4 (Nahant Update 9) if
> that makes any difference.
>
> Can anybody tell me what might be the problem or how to fix this issue.
>
> thanks
> Mike
>
--
Pavan Balaji
http://www.mcs.anl.gov/~balaji
More information about the mpich-discuss
mailing list