This appears to be a legit bug with the use of MPI_T in the test/example monitoring app, so I'm going to move the discussion to the Github issue so that we can track it properly:
https://github.com/open-mpi/ompi/issues/9260 To answer Jong's question: ob1 is one of Open MPI's point-to-point messaging plugins. PML plugins are basically the back-end of MPI_Send, MPI_Recv, and other point-to-point MPI APIs. See https://github.com/open-mpi/ompi/blob/v4.1.x/README#L697-L826 and slides 33-42 in https://www.open-mpi.org/video/general/easybuild_tech_talks_01_OpenMPI_part2_20200708.pdf (see the all 3 presentation videos and PDFs of the slides here: https://www.open-mpi.org/video/?category=general). On Aug 17, 2021, at 5:06 PM, Jong Choi via users <users@lists.open-mpi.org> wrote: Thank you for the information. I don't know what ob1 is and possible other choices are. Is there any way for me to check? Anyhow, I tried a few things but got the same error. Here ia bit more verbose output: shell$ mpirun -n 1 --allow-run-as-root --mca pml_base_verbose 10 --mca mtl_base_verbose 10 --mca osc_base_verbose 10 --mca pml ob1,monitoring --mca pml_monitoring_enable 2 ./check_monitoring [32c7ade48bf4:52151] mca: base: components_register: registering framework pml components [32c7ade48bf4:52151] mca: base: components_register: found loaded component ob1 [32c7ade48bf4:52151] mca: base: components_register: component ob1 register function successful [32c7ade48bf4:52151] mca: base: components_register: found loaded component monitoring [32c7ade48bf4:52151] mca: base: components_register: component monitoring register function successful [32c7ade48bf4:52151] mca: base: components_open: opening pml components [32c7ade48bf4:52151] mca: base: components_open: found loaded component ob1 [32c7ade48bf4:52151] mca: base: components_open: component ob1 open function successful [32c7ade48bf4:52151] mca: base: components_open: found loaded component monitoring [32c7ade48bf4:52151] mca: base: components_open: component monitoring open function successful [32c7ade48bf4:52151] mca: base: components_register: registering framework osc components [32c7ade48bf4:52151] mca: base: components_register: found loaded component pt2pt [32c7ade48bf4:52151] mca: base: components_register: component pt2pt register function successful [32c7ade48bf4:52151] mca: base: components_register: found loaded component monitoring [32c7ade48bf4:52151] mca: base: components_register: component monitoring register function successful [32c7ade48bf4:52151] mca: base: components_register: found loaded component rdma [32c7ade48bf4:52151] mca: base: components_register: component rdma register function successful [32c7ade48bf4:52151] mca: base: components_register: found loaded component sm [32c7ade48bf4:52151] mca: base: components_register: component sm register function successful [32c7ade48bf4:52151] mca: base: components_open: opening osc components [32c7ade48bf4:52151] mca: base: components_open: found loaded component pt2pt [32c7ade48bf4:52151] mca: base: components_open: found loaded component monitoring [32c7ade48bf4:52151] mca: base: components_open: found loaded component rdma [32c7ade48bf4:52151] mca: base: components_open: found loaded component sm [32c7ade48bf4:52151] mca: base: components_open: component sm open function successful [32c7ade48bf4:52151] select: initializing pml component ob1 [32c7ade48bf4:52151] select: init returned priority 20 [32c7ade48bf4:52151] select: initializing pml component monitoring [32c7ade48bf4:52151] select: init returned priority 0 [32c7ade48bf4:52151] selected ob1 best priority 20 [32c7ade48bf4:52151] select: component ob1 selected [32c7ade48bf4:52151] select: component monitoring not selected / finalized [32c7ade48bf4:52151] mca: base: close: component monitoring closed [32c7ade48bf4:52151] mca: base: close: unloading component monitoring [32c7ade48bf4:52151] check:select: PML check not necessary on self [32c7ade48bf4:52151] mca: base: components_register: registering framework mtl components Cannot find monitoring MPI_Tool "pml_monitoring_messages_count" pvar, check that you have enabled the monitoring component. -- Jeff Squyres jsquy...@cisco.com
