Wire pali_chaos_kek.js to run automatically in Evergreen

    • Type: Task
    • Resolution: Unresolved
    • Priority: Major - P3
    • None
    • Affects Version/s: None
    • Component/s: None
    • None
    • Server Security
    • None
    • None
    • None
    • None
    • None
    • None
    • None

      The SERVER-129742 pali_chaos_kek.js Jira ticket does not schedule the test in Evergreen. It needs its own resmoke suite file and a corresponding Evergreen task entry, following the same pattern as disagg_pali_chaos.

      Work Items

      • Create buildscripts/modules/atlas/suites/disagg_pali_chaos_kek.yml. }}Copy {{disagg_pali_chaos.yml verbatim and change the selector root:
      test_kind: js_testselector:
        roots:
          - src/mongo/db/modules/*/jstests/pali_chaos/pali_chaos_kek.jsexecutor:
        config:
          shell_options:
            nodb: ""
            global_vars:
              TestData:
                doesNotSupportGracefulStepdown: true
                skipCollectionAndIndexValidation: true
        hooks:
          - class: ContainerCleanup
        archive:
          tests: true
          hooks: {} 
      • Add a task entry in etc/evergreen_yml_components/tasks/resmoke/server_divisions/durable_transactions_and_availability/tasks.yml. Add immediately after the disagg_pali_chaos task block (line 2752):
        - <<: *task_template
          name: disagg_pali_chaos_kek
          tags:
            [
              "assigned_to_jira_team_server_storage_engines",
              "default",
              "incompatible_tsan",
              "large",
              "requires_large_host",
              "requires_extra_system_deps",
            ]
          commands:
            - func: "do setup"
            - func: "run tests"
              vars:
                suite: disagg_pali_chaos_kek
                resmoke_jobs_max: 1 

      Rationale for tag choices

      All tags are copied from disagg_pali_chaos:

      • requires_extra_system_deps — the test needs Docker and PyKMIP; this tag excludes it from the standard dev waterfall variants that run without extra infrastructure
      • requires_large_host — SLSMinimalThreeCellTest launches ~15 Docker containers; running on a small host exhausts Docker bridge networking
      • resmoke_jobs_max: 1 — same reason: multiple concurrent SLS-heavy tests on the same host exhaust port mappings
      • incompatible_tsan — SLS native libraries are not TSAN-instrumented default — schedules it in the same variants that already run disagg_pali_chaos

      Expected outcome

      After merging, the disagg_pali_chaos_kek task appears in Evergreen alongside disagg_pali_chaos in whatever variants currently schedule that task. No changes to variant YAML are required — the tag-based selection handles it automatically.

            Assignee:
            Unassigned
            Reporter:
            Chye Lin Chee
            Votes:
            0 Vote for this issue
            Watchers:
            1 Start watching this issue

              Created:
              Updated: