PMIX ERR NOMEM error, but the test can still run
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 203
- Forks
- 67
- Avg merge
- 20m
- Merged PRs (30d)
- 8
Description
Hi
I am running mlpstorage training run test, but I saw some PMIX ERR NOMEM error during the test, but it seems the test is not stopped and it is still running, and at the end I can see a success/fail test result.
Is this PMIX ERR NOMEM error a big concern and the test result is good to be accepted ? pls help to comment, thanks.
//log
(myenv) root@minio1:~# mlpstorage training run --hosts 10.17.73.100 10.17.73.103 10.17.73.106 10.17.73.107 10.17.73.108 10.17.73.109 10.17.73.111 10.17.73.112 --client-host-memory-in-gb 64 --num-accelerators 132 --accelerator-type a100 --model resnet50 --data-dir /mnt/newdisk/fcresnet50/ --results-dir /home/test/resnet50_results/ --param checkpoint.checkpoint_folder=/home/test/ckpoint --param dataset.num_files_train=21103 --param reader.computation_threads=48 --param reader.prefetch_size=0 --param reader.transfer_size=768 --mpi-params "--mca btl_tcp_if_include ens36" --allow-run-as-root --closed
Setting attr from num_accelerators to 132
Hosts is: ['10.17.73.100', '10.17.73.103', '10.17.73.106', '10.17.73.107', '10.17.73.108', '10.17.73.109', '10.17.73.111', '10.17.73.112']
Hosts is: ['10.17.73.100', '10.17.73.103', '10.17.73.106', '10.17.73.107', '10.17.73.108', '10.17.73.109', '10.17.73.111', '10.17.73.112']
2025-08-24 15:39:17|STATUS: Benchmark results directory: /home/test/resnet50_results/training/resnet50/run/20250824_153917
2025-08-24 15:39:17|INFO: Found benchmark run: training_run_resnet50_20250824_153917
2025-08-24 15:39:17|STATUS: Verifying benchmark run for training_run_resnet50_20250824_153917
2025-08-24 15:39:17|RESULT: Minimum file count dictated by 500 step requirement of given accelerator count and batch size.
2025-08-24 15:39:17|STATUS: Closed: [CLOSED] Closed parameter override allowed: checkpoint.checkpoint_folder = /home/test/ckpoint (Parameter: Overrode Parameters)
2025-08-24 15:39:17|STATUS: Closed: [CLOSED] Closed parameter override allowed: dataset.num_files_train = 21103 (Parameter: Overrode Parameters)
2025-08-24 15:39:17|STATUS: Closed: [CLOSED] Closed parameter override allowed: reader.computation_threads = 48 (Parameter: Overrode Parameters)
2025-08-24 15:39:17|STATUS: Closed: [CLOSED] Closed parameter override allowed: reader.prefetch_size = 0 (Parameter: Overrode Parameters)
2025-08-24 15:39:17|STATUS: Closed: [CLOSED] Closed parameter override allowed: reader.transfer_size = 768 (Parameter: Overrode Parameters)
2025-08-24 15:39:17|STATUS: Benchmark run qualifies for CLOSED category ([RunID(program='training', command='run', model='resnet50', run_datetime='20250824_153917')])
2025-08-24 15:39:17|STATUS: Running benchmark command:: mpirun -n 132 -host 10.17.73.100:17,10.17.73.103:17,10.17.73.106:17,10.17.73.107:17,10.17.73.108:16,10.17.73.109:16,10.17.73.111:16,10.17.73.112:16 --allow-run-as-root --mca btl_tcp_if_include ens36 /root/.venvs/myenv/bin/dlio_benchmark workload=resnet50_a100 ++hydra.run.dir=/home/test/resnet50_results/training/resnet50/run/20250824_153917 ++hydra.output_subdir=dlio_config ++workload.checkpoint.checkpoint_folder=/home/test/ckpoint ++workload.dataset.num_files_train=21103 ++workload.reader.computation_threads=48 ++workload.reader.prefetch_size=0 ++workload.reader.transfer_size=768 ++workload.dataset.data_folder=/mnt/newdisk/fcresnet50/resnet50 --config-dir=/root/storage/configs/dlio
[minio1:768540] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../../src/mca/gds/shmem/gds_shmem.c at line 1958
[minio1:768540] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../../src/mca/gds/shmem/gds_shmem.c at line 1966
[minio1:768540] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../src/mca/gds/base/gds_base_fns.c at line 238
[minio1:768540] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../src/mca/gds/base/gds_base_fns.c at line 255
[minio1:768540] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../src/server/pmix_server.c at line 3409
[minio1:220609] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../../src/mca/gds/shmem/gds_shmem.c at line 1958
[minio1:220609] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../../src/mca/gds/shmem/gds_shmem.c at line 1966
[minio1:220609] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../src/mca/gds/base/gds_base_fns.c at line 238
[minio1:220609] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../src/mca/gds/base/gds_base_fns.c at line 255
[minio1:220609] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../src/server/pmix_server.c at line 3409
[minio1:224733] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../../src/mca/gds/shmem/gds_shmem.c at line 1958
[minio1:224733] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../../src/mca/gds/shmem/gds_shmem.c at line 1966
[minio1:224733] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../src/mca/gds/base/gds_base_fns.c at line 238
[minio1:224733] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../src/mca/gds/base/gds_base_fns.c at line 255
[minio1:224733] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../src/server/pmix_server.c at line 3409
[minio1:226590] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../../src/mca/gds/shmem/gds_shmem.c at line 1958
[minio1:226590] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../../src/mca/gds/shmem/gds_shmem.c at line 1966
[minio1:226590] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../src/mca/gds/base/gds_base_fns.c at line 238
[minio1:226590] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../src/mca/gds/base/gds_base_fns.c at line 255
[minio1:226590] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../src/server/pmix_server.c at line 3409
[minio1:228063] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../../src/mca/gds/shmem/gds_shmem.c at line 1958
[minio1:228063] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../../src/mca/gds/shmem/gds_shmem.c at line 1966
[minio1:228063] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../src/mca/gds/base/gds_base_fns.c at line 238
[minio1:228063] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../src/mca/gds/base/gds_base_fns.c at line 255
[minio1:228063] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../src/server/pmix_server.c at line 3409
[minio1:230458] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../../src/mca/gds/shmem/gds_shmem.c at line 1958
[minio1:230458] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../../src/mca/gds/shmem/gds_shmem.c at line 1966
[minio1:230458] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../src/mca/gds/base/gds_base_fns.c at line 238
[minio1:230458] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../src/mca/gds/base/gds_base_fns.c at line 255
[minio1:230458] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../src/server/pmix_server.c at line 3409
[minio1:228468] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../../src/mca/gds/shmem/gds_shmem.c at line 1958
[minio1:228468] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../../src/mca/gds/shmem/gds_shmem.c at line 1966
[minio1:228468] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../src/mca/gds/base/gds_base_fns.c at line 238
[minio1:228468] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../src/mca/gds/base/gds_base_fns.c at line 255
[minio1:228468] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../src/server/pmix_server.c at line 3409
[minio1:226537] PMIX ERROR: PMIX_ERR_NOMEM in file ../../../../../../src/mca/gds/shmem/gds_shmem.c at line 1958
Thanks,
Mike Chen
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Reproduce the mlpstorage training run command shown in the issue across the listed hosts and capture the PMIX log alongside the final benchmark result. The payload names no source file or test; trace the command's MPI/PMIX execution path and determine whether PMIX_ERR_NOMEM affects result validity, then add a reproducible check or documented handling.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- distributed-systems, performance
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100