Disagg python testing: initial triage of (previously) failing tests

XMLWordPrintableJSON

    • Type: Task
    • Resolution: Fixed
    • Priority: Major - P3
    • disagg-temp
    • Affects Version/s: None
    • Component/s: Test Python
    • None
    • Storage Engines, Storage Engines - Foundations
    • SE Foundations - 2025-06-20
    • 2

      In January, SLS-1040 created test/format/hook_disagg.fail, listing all the python test files that had failure (or ran too long) when running with the disagg tests.  There are 250 of them.  Just now, I reran this list without any time limits, and the vast majority succeed.

      Here's the recipe for checking the failures:

      sed -e 's/ *#.*//' -e '/^$/d' <  test/suite/hook_disagg.fail > hook.run
      cd build
      ../tools/pytest_parallel -j 4 -v 2 -- `cat ../hook.run` > ../parallel.out 2>&1 & 

      The current output gives a summary of failures:

      Summary:
        ERROR: test_cc03.py process returned 1
        ERROR: test_checkpoint08.py process returned 1
        ERROR: test_checkpoint10.py process returned 134
        ERROR: test_checkpoint11.py process returned 134
        ERROR: test_checkpoint15.py process returned 134
        ERROR: test_checkpoint14.py process returned 134
        ERROR: test_checkpoint20.py process returned 134
        ERROR: test_checkpoint21.py process returned 134
        ERROR: test_checkpoint18.py process returned 134
        ERROR: test_checkpoint25.py process returned 134
        ERROR: test_checkpoint28.py process returned 134
        ERROR: test_cursor10.py process returned 1
        ERROR: test_dump.py process returned 134
        ERROR: test_rename.py process returned 1
        ERROR: test_upgrade.py process returned 1
        ERROR: test_util16.py process returned 1
        ERROR: test_util20.py process returned 1
      

      The work in this ticket is to update hook_disagg.fail to remove the successful files, and verify that they work reliably. All tests (except from hook_disagg.fail) run as part of the pull request for PRs to the disagg branch.

            Assignee:
            Donald Anderson
            Reporter:
            Donald Anderson
            Votes:
            0 Vote for this issue
            Watchers:
            1 Start watching this issue

              Created:
              Updated:
              Resolved: