Andrej, that really looks like a SLURM issue that does not involve Open MPI
In order to confirm, you can $ salloc -N 2 -n 2 /* and then from the allocation */ srun hostname If this does not work, then this is a SLURM issue you have to fix. Once fixed, I am confident Open MPI will just work Cheers, Gilles On Tue, Feb 2, 2021 at 10:22 AM Andrej Prsa via devel <devel@lists.open-mpi.org> wrote: > > Hi Gilles, > > > Here is what you can try > > > > $ salloc -N 4 -n 384 > > /* and then from the allocation */ > > > > $ srun -n 1 orted > > /* that should fail, but the error message can be helpful */ > > > > $ /usr/local/bin/mpirun --mca plm slurm --mca plm_base_verbose 10 true > > andrej@terra:~/system/tests/MPI$ salloc -N 4 -n 384 > salloc: Granted job allocation 837 > andrej@terra:~/system/tests/MPI$ srun -n 1 orted > srun: Warning: can't run 1 processes on 4 nodes, setting nnodes to 1 > srun: launch/slurm: launch_p_step_launch: StepId=837.0 aborted before > step completely launched. > srun: Job step aborted: Waiting up to 32 seconds for job step to finish. > srun: error: task 0 launch failed: Unspecified error > andrej@terra:~/system/tests/MPI$ /usr/local/bin/mpirun -mca plm slurm > -mca plm_base_verbose 10 true > [terra:179991] mca: base: components_register: registering framework plm > components > [terra:179991] mca: base: components_register: found loaded component slurm > [terra:179991] mca: base: components_register: component slurm register > function successful > [terra:179991] mca: base: components_open: opening plm components > [terra:179991] mca: base: components_open: found loaded component slurm > [terra:179991] mca: base: components_open: component slurm open function > successful > [terra:179991] mca:base:select: Auto-selecting plm components > [terra:179991] mca:base:select:( plm) Querying component [slurm] > [terra:179991] [[INVALID],INVALID] plm:slurm: available for selection > [terra:179991] mca:base:select:( plm) Query of component [slurm] set > priority to 75 > [terra:179991] mca:base:select:( plm) Selected component [slurm] > [terra:179991] plm:base:set_hnp_name: initial bias 179991 nodename hash > 2928217987 > [terra:179991] plm:base:set_hnp_name: final jobfam 7711 > [terra:179991] [[7711,0],0] plm:base:receive start comm > [terra:179991] [[7711,0],0] plm:base:setup_job > [terra:179991] [[7711,0],0] plm:slurm: LAUNCH DAEMONS CALLED > [terra:179991] [[7711,0],0] plm:base:setup_vm > [terra:179991] [[7711,0],0] plm:base:setup_vm creating map > [terra:179991] [[7711,0],0] plm:base:setup_vm add new daemon [[7711,0],1] > [terra:179991] [[7711,0],0] plm:base:setup_vm assigning new daemon > [[7711,0],1] to node node9 > [terra:179991] [[7711,0],0] plm:base:setup_vm add new daemon [[7711,0],2] > [terra:179991] [[7711,0],0] plm:base:setup_vm assigning new daemon > [[7711,0],2] to node node10 > [terra:179991] [[7711,0],0] plm:base:setup_vm add new daemon [[7711,0],3] > [terra:179991] [[7711,0],0] plm:base:setup_vm assigning new daemon > [[7711,0],3] to node node11 > [terra:179991] [[7711,0],0] plm:base:setup_vm add new daemon [[7711,0],4] > [terra:179991] [[7711,0],0] plm:base:setup_vm assigning new daemon > [[7711,0],4] to node node12 > [terra:179991] [[7711,0],0] plm:slurm: launching on nodes > node9,node10,node11,node12 > [terra:179991] [[7711,0],0] plm:slurm: Set prefix:/usr/local > [terra:179991] [[7711,0],0] plm:slurm: final top-level argv: > srun --ntasks-per-node=1 --kill-on-bad-exit --ntasks=4 orted -mca > ess "slurm" -mca ess_base_jobid "505348096" -mca ess_base_vpid "1" -mca > ess_base_num_procs "5" -mca orte_node_regex > "terra,node[1:9],node[2:10-12]@0(5)" -mca orte_hnp_uri > "505348096.0;tcp://10.9.2.10,192.168.1.1:38995" -mca plm_base_verbose "10" > [terra:179991] [[7711,0],0] plm:slurm: reset PATH: > /usr/local/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin > [terra:179991] [[7711,0],0] plm:slurm: reset LD_LIBRARY_PATH: /usr/local/lib > srun: launch/slurm: launch_p_step_launch: StepId=837.1 aborted before > step completely launched. > srun: Job step aborted: Waiting up to 32 seconds for job step to finish. > srun: error: task 3 launch failed: Unspecified error > srun: error: task 1 launch failed: Unspecified error > srun: error: task 2 launch failed: Unspecified error > srun: error: task 0 launch failed: Unspecified error > [terra:179991] [[7711,0],0] plm:slurm: primary daemons complete! > [terra:179991] [[7711,0],0] plm:base:receive stop comm > [terra:179991] mca: base: close: component slurm closed > [terra:179991] mca: base: close: unloading component slurm > > This is what I'm seeing in slurmctld.log: > > [2021-02-01T20:15:18.358] sched: _slurm_rpc_allocate_resources JobId=837 > NodeList=node[9-12] usec=537 > [2021-02-01T20:15:26.815] error: mpi_hook_slurmstepd_prefork failure for > 0x557ce5b92960s on node9 > [2021-02-01T20:15:59.621] error: mpi_hook_slurmstepd_prefork failure for > 0x55cc6c89a7e0s on node12 > [2021-02-01T20:15:59.621] error: mpi_hook_slurmstepd_prefork failure for > 0x55b7b8b467e0s on node10 > [2021-02-01T20:15:59.622] error: mpi_hook_slurmstepd_prefork failure for > 0x55f8cd69a7e0s on node11 > [2021-02-01T20:15:59.628] error: mpi_hook_slurmstepd_prefork failure for > 0x5555b45bc7e0s on node9 > > And this is in slurmd.node9.log (and similar for the remaining 3 nodes): > > [2021-02-01T20:15:59.592] task/affinity: lllp_distribution: JobId=837 > manual binding: none > [2021-02-01T20:15:59.624] [837.1] error: node9 [0] pmixp_client_v2.c:246 > [pmixp_lib_init] mpi/pmix: ERROR: PMIx_server_init failed with error -2 > : Success (0) > [2021-02-01T20:15:59.624] [837.1] error: node9 [0] pmixp_client.c:518 > [pmixp_libpmix_init] mpi/pmix: ERROR: PMIx_server_init failed with error -1 > : Success (0) > [2021-02-01T20:15:59.624] [837.1] error: node9 [0] pmixp_server.c:423 > [pmixp_stepd_init] mpi/pmix: ERROR: pmixp_libpmix_init() failed > [2021-02-01T20:15:59.624] [837.1] error: (null) [0] mpi_pmix.c:169 > [p_mpi_hook_slurmstepd_prefork] mpi/pmix: ERROR: pmixp_stepd_init() failed > [2021-02-01T20:15:59.627] [837.1] error: Failed mpi_hook_slurmstepd_prefork > [2021-02-01T20:15:59.650] [837.1] error: job_manager exiting abnormally, > rc = -1 > [2021-02-01T20:16:02.000] [837.1] done with job > > Cheers, > Andrej >