[mpich-discuss] gdb + hydra

Christoph Sprenger csprenger at wetafx.co.nz
Tue Sep 18 01:04:12 CDT 2012


still trying to get this to work.
i've also tried a vanilla tcsh and bash with no env set at all and both 
seem to fail still.

same things persists when running mpiexec -n 1 gdb without xterm

Ubuntu 10.04.4 LTS
GNU bash, version 4.1.5(1)-release
tcsh 6.17.00 (Astron)


mpich2
./configure --enable-shared --enable-sharedlibs=gcc




#include "mpi.h"
#include <iostream>

int main(int argc, char *argv[])
{
     MPI_Init(&argc, &argv);
     int rank,size;
     MPI_Comm_size( MPI_COMM_WORLD, &size );
     MPI_Comm_rank( MPI_COMM_WORLD, &rank );
     std::cerr << "Hello World from rank " << rank << " / " << size << 
std::endl;
     MPI_Finalize();
}

[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/libthread_db.so.1".
[cli_0]: write_line error; fd=6 buf=:cmd=init pmi_version=1 pmi_subversion=1
:
system msg for write_line failure : Bad file descriptor
[cli_0]: Unable to write to PMI_fd
[cli_0]: write_line error; fd=6 buf=:cmd=get_appnum
:
system msg for write_line failure : Bad file descriptor
Fatal error in MPI_Init: Other MPI error, error stack:
MPIR_Init_thread(392):
MPID_Init(107).......: channel initialization failed
MPID_Init(389).......: PMI_Get_appnum returned -1
[Inferior 1 (process 3132) exited with code 01]


is there anything else i can try to make this work ( eg different 
configure options for hydra or mpich2 ) ?

Cheers,
Christoph




On 17/09/12 20:00, Christoph Sprenger wrote:
> I'm having some issues doing the following:
>
> mpiexec -n 1 xterm -e gdb -ex r --args ./app ...
>
>
> this used to work fine for me with 1.2.1 and mpd but seems to give me :
>
> [cli_0]: write_line error; fd=6 buf=:cmd=init pmi_version=1 
> pmi_subversion=1
> :
> system msg for write_line failure : Bad file descriptor
> [cli_0]: Unable to write to PMI_fd
> [cli_0]: write_line error; fd=6 buf=:cmd=get_appnum
>
> also tried :
> setenv HYDRA_LAUNCHER_AUTOFORK 0
>
> however it falls over in MPI_Init.
> seems like other people experienced this too, but i couldn't find a 
> hint that seems to address this issue
>
> i've tried tried 1.4.1p1 and 1.5rc1.
>
> any help would be appreciated.
>
> Kind Regards,
> Christoph
>
> _______________________________________________
> mpich-discuss mailing list     mpich-discuss at mcs.anl.gov
> To manage subscription options or unsubscribe:
> https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss



More information about the mpich-discuss mailing list