We are currently compiling a list of problems and their likely solutions:
Site |
Error |
MPI version |
Proposed Solution |
Solution Verified by site |
argoce01.na.infn.it |
Unable to verify signature! Server |
OpenMPI |
CA cert not installed |
|
argoce01.na.infn.it |
sensors/testjob-mpi/.[a-zA-Z0-9]*: Warning: Cannot stat: No such file or directory |
OpenMPI |
Unable to copy data to WN |
|
ce-01.grid.sissa.it |
mpicc not found |
OpenMPI |
openmpi-devel RPM should be installed |
Reconfiguration of variables by site. Site uses modified mpi-start to load modules |
ce-cyb.ca.infn.it |
failed to find requested MPI type : mpich |
mpich 1.2.7p7 |
environment on Workernode may not be set up correctly |
|
ce.bfg.uni-freiburg.de |
mpicc not found |
openmpi 1.1.1 |
openmpi-devel not installed? or path /usr not correct |
|
ce.cyb-pcr.it |
job proxy expired - no tags published (verify) |
|
Job did not get allocated nodes |
|
ce.grid.rug.nl |
no mpiexec found in ... /opt/i2g/bin/../etc/mpi-start/generic_mpiexec.sh: line 23: -machinefile: command not found |
mpich |
environment not set up correctly |
|
ce.grid.rug.nl |
mpiexec noticed that job rank 0 with PID 18646 on node node6 exited on signal 11 (Segmentation fault). |
openmpi |
??? |
|
ce.ngcc.acad.bg |
mpicc not found |
mpich 1.2.7 |
mpich-devel RPM should be installed |
|
ce.reef.man.poznan.pl |
tar: /tmp/1447602.ce.reef.man.poznan.pl/https_3a_2f_2fwms220.cern.ch_3a9000_2f5mPVR9Zm-5xTnFGQNDygTA/sensors/testjob-mpi/.[a-zA-Z0-9]*: Warning: Cannot stat: No such file or directory |
mpich2 |
unshared homes file distribution not working? |
|
ce.reef.man.poznan.pl |
As above, but Fatal error in MPI_Send: |
mpich2 |
environment not set up correctly (uses mvapich) |
|
ce01-lhcb-t2.cr.cnaf.infn.it |
/usr/bin/ld: cannot find -lmpich |
mpich 1.2.7 |
RPM not correctly installed (mpicc is ok) |
|
ce01.ariagni.hellasgrid.gr |
/opt/mpiexec-0.82/bin/mpiexec: error while loading shared libraries: libtorque.so.0: cannot open shared object file: No such file or directory |
mpich |
mpiexec version is wrong for installed version of torque |
|
ce01.ariagni.hellasgrid.gr |
MPI_SPECIFIC_PARAMS+=-x X509_USER_PROXY --prefix /opt/openmpi/1.1 : No such file or directory |
openmpi, mpich2 |
environment incorrect ?? |
|
ce01.athena.hellasgrid.gr |
/opt/i2g/bin/../etc/mpi-start/openmpi.mpi: line 70: MPI_SPECIFIC_PARAMS+=-x X509_USER_PROXY --prefix /opt/openmpi-hg-1.3.3-gcc/64 |
openmpi |
environment incorrect |
|
ce01.grid.info.uvt.ro |
I2G_MPI_START variable is not set! |
mpi |
environment incorrect |
|
ce01.isabella.grnet.gr |
/opt/i2g/bin/../etc/mpi-start/openmpi.mpi: line 70: MPI_SPECIFIC_PARAMS+=-x X509_USER_PROXY --prefix / : No such file or directory |
openmpi |
MPI_OPENMPI_PATH=/ does not set the environment correctly |
|
grid-ce01.esrf.eu |
pingtest.c:2:17: mpi.h: No such file or directory |
mpich2 |
Likely that mpich2-devel not installed |
|
kg-ce01.cc.kuleuven.be |
openmpi compile failure due to incompatible libraries |
openmpi (64 bit) |
due to site using MPI_MPICC_OPTS="-m32" with 64bit openmpi |
|
TCG working