Simulations running error [Reason=user_env_retrieval_failed_requeued_held] on the HPC #32402
-
Check these boxes if you have followed the posting rules.
QuestionDear all, After the maintance of the HPC in our university, the PorousFLow module did not work (“Reason=user_env_retrieval_failed_requeued_held”) after the submission of job. I got the follwing suggestions from the HPC engineer of my university. _"something seems to be wrong/misconfigured with your login procedure. Could you please remove the conda part from your .bashrc or fix your conda environment?"_ I tried to uncomment the "conda activate moose" and "conda initialize" from my .bashrc, then I can run the simulation after submit the task. But I got the error "mpirun: command not found..." when I log out and then login into the HPC. Could you please help me on this ? Thanks a lot. Best regards. Appendix 1: Jobscript fot the job: #!/bin/bash #module load ${MPI_MODULE} cd ${SLURM_SUBMIT_DIR} echo $startexe exit Appendix 2: Job detail: JobId=3307221 JobName=job.sh |
Beta Was this translation helpful? Give feedback.
Replies: 2 comments 21 replies
-
|
Hello We do not recommend using conda to install the code on HPC as the binaries would not be linked against the local MPI libraries that are installed by the cluster administrators. Please re install moose using the HPC instructions instead. |
Beta Was this translation helpful? Give feedback.
-
|
Thanks for your suggestions, Guillaume. Actually, a few years ago I followed the MOOSE group's recommendation to use conda for installation on the HPC, after encountering issues with the HPC instructions. This time, I'm trying the HPC installation instructions again. Thanks again for your help. Best, |
Beta Was this translation helpful? Give feedback.
-
|
Hello @GiudGiud and other experts, I follow the HPC instructions. After finishing PETSc, libMesh, and WASP without errors, the Build and Test MOOSE produce the following error information. Could you please help me on this? Thanks a lot. Best, (base) [ka_qz9211@uc3n991 test]$ make -j 6 Traceback (most recent call last): Traceback (most recent call last): ERROR: Your Plugin Tester "CSVValidationTester" failed to import. (skipping) No module named 'numpy' ERROR: Your Plugin Tester "XMLDiff" failed to import. (skipping) No module named 'numpy' ERROR: Your Plugin Tester "AnalyzeJacobian" failed to import. (skipping) No module named 'numpy' ERROR: Your Plugin Tester "MMSTest" failed to import. (skipping) No module named 'numpy' ERROR: Your Plugin Tester "JSONDiff" failed to import. (skipping) No module named 'numpy' ERROR: Your Plugin Tester "RunCommand" failed to import. (skipping) No module named 'numpy' ERROR: Your Plugin Tester "ImageDiff" failed to import. (skipping) No module named 'numpy' ERROR: Your Plugin Tester "RunApp" failed to import. (skipping) No module named 'numpy' ERROR: Your Plugin Tester "PythonUnitTest" failed to import. (skipping) No module named 'numpy' ERROR: Your Plugin Tester "FileTester" failed to import. (skipping) No module named 'numpy' ERROR: Your Plugin Tester "SchemaDiff" failed to import. (skipping) No module named 'numpy' ERROR: Your Plugin Tester "PetscJacobianTester" failed to import. (skipping) No module named 'numpy' ERROR: Your Plugin Tester "RunException" failed to import. (skipping) No module named 'numpy' ERROR: Your Plugin Tester "SignalTester" failed to import. (skipping) No module named 'numpy' ERROR: Your Plugin Tester "Tester" failed to import. (skipping) No module named 'numpy' ERROR: Your Plugin Tester "CheckFiles" failed to import. (skipping) No module named 'numpy' ERROR: Your Plugin Tester "Exodiff" failed to import. (skipping) No module named 'numpy' ERROR: Your Plugin Tester "CSVDiff" failed to import. (skipping) No module named 'numpy' ################################################################################ [0]PETSC ERROR: --------------------- Error Message -------------------------------------------------------------- 2.85 seconds |
Beta Was this translation helpful? Give feedback.
-
|
Thanks for your suggestions. I delete compiled objects and MOOSE folders to restart the installation. I try the all the options of MPI "mpi/impi/2021.14-intel-2025.0 For example, I set "module load mpi/openmpi/5.0-gnu-14.2 devel/python/3.12.3-gnu-14.2" at the beginning. However, all options fail with same errors when I set "./update_and_rebuild_petsc.sh || return". But the "./diagnostics.sh" shows no errors. |
Beta Was this translation helpful? Give feedback.
-
Seems to be trying to build with cuda support, but failing to do so. If not try to clear your environment variables (PATH, CXX_INCLUDE_DIR, LD_LIBRARY_PATH) of these so the cuda content is not found |
Beta Was this translation helpful? Give feedback.
-
|
Could please let me know how to clear theenvironment variables (PATH, CXX_INCLUDE_DIR, LD_LIBRARY_PATH) of these ? by setting "module unload devel/cuda/12.8" ? |
Beta Was this translation helpful? Give feedback.
-
|
you can re-export environment variable |
Beta Was this translation helpful? Give feedback.
you can re-export environment variable