HI Yann, the orte_run message (Error obtaining unique transport key from ORTE (orte_precondition_transports not present in the environment) is normal with OpenMPI on QLogic/Intel/PSM fabrics.
There are two ways to avoid it: 1) disable PSM (obviously not recommended for parallel jobs): export OMPI_MCA_mtl=^psm 2) set the transport key manually: eval "export $(mpiexec env | grep OMPI_MCA_orte_precondition_transports | head -n 1)" see also here: https://www.open-mpi.org/community/lists/users/2011/10/17660.php Regards, Bart On 30 June 2016 at 04:17, Yann Sagon <yann.sa...@unige.ch> wrote: > Hello, > > After some investigation, the error is coming because when the compilation > of Rmpi is done. The script try to load the module to see if it's working. > It seems it's loading without mpirun or orterun or similar and thus fail > with orte_init. > > I have tried to load the compiled module when launching R with mpirun and > it's working fine. > > I didn't wanted to update the non standard exts version, I just wanted to > have a newer version of R. > > Anyway, I have no idea why the installation of Rmpi was working with > previous version of R. > > In the last line of this file, the comment seems to suggest that it might > be needed to launch the compilation with mpirun but I don't know how to do > that with eb. > > > https://github.com/hpcugent/easybuild-easyblocks/blob/master/easybuild/easyblocks/r/rmpi.py > > When compiling manually like that, it'w working: > > mpirun -np 1 R CMD INSTALL > /home/sagon/.local/easybuild/sources/r/R/extensions/Rmpi_0.6-6.tar.gz > > --configure-args="--with-Rmpi-include=/opt/ebsofts/Compiler/GCC/4.9.3-2.25/OpenMPI/1.10.2/include > --with-Rmpi-libpath=/opt/ebsofts/Compiler/GCC/4.9.3-2.25/OpenMPI/1.10.2/lib > --with-mpi=/opt/ebsofts/Compiler/GCC/4.9.3-2.25/OpenMPI/1.10.2 > --with-Rmpi-type=OPENMPI" > > --library=/opt/ebsofts/MPI/GCC/4.9.3-2.25/OpenMPI/1.10.2/R/3.3.1/lib64/R/library > --no-clean-on-error > > without mpirun, it' failed with the test. > > > 2016-06-29 18:00 GMT+02:00 Kenneth Hoste <kenneth.ho...@ugent.be>: > >> Hi Yann, >> >> On 29/06/16 17:51, Yann Sagon wrote: >> > Hello, >> > >> > I'm trying to tweak the file >> > >> > R-3.2.3-foss-2016a.eb >> > >> > to install version 3.3.1 of R. >> > >> > In this file, there is a list of non standard exts. >> > >> > Rmpi seems to do some checking once installed and failed with orte_init. >> > >> > Any clue? >> >> Please share the full error message you're running into? >> >> There is an ongoing effort to add an easyconfig for R 3.3.x, but it >> seems like it's stuck in dep hell, see >> https://github.com/hpcugent/easybuild-easyconfigs/pull/3079 . >> >> >> regards, >> >> Kenneth >> > > > > -- > Yann SAGON > Ingénieur système HPC > 24 Rue du Général-Dufour > 1211 Genève 4 - Suisse > Tél. : +41 (0)22 379 7737 > yann.sa...@unige.ch - www.unige.ch > -- Dr. Bart E. Oldeman | bart.olde...@mcgill.ca | bart.olde...@calculquebec.ca Scientific Computing Analyst / Analyste en calcul scientifique McGill HPC Centre / Centre de Calcul Haute Performance de McGill | http://www.hpc.mcgill.ca Calcul Québec | http://www.calculquebec.ca Compute/Calcul Canada | http://www.computecanada.ca Tel/Tél: 514-396-8926 | Fax/Télécopieur: 514-396-8934
