Uploaded image for project: 'Lustre'
  1. Lustre
  2. LU-21223

sanity-quota test_84: dd succeed, expect EDQUOT after quota reset on ZFS

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Unresolved
    • Icon: Medium Medium
    • None
    • None
    • None
    • 3
    • 9223372036854775807

      This issue was created by maloo for Patrickbot <patrick-bot@mulberrytree.us>

      This issue relates to the following test suite run: https://testing.whamcloud.com/test_sets/47ccb8eb-048d-4b8b-980c-7a8680af3177

      test_84 failed with the following error:

      dd succeed, expect EDQUOT
      

      Test session details:

      Client
      Lustre: 2.17.61.87
      Branch: master
      Build: https://build.whamcloud.com/job/lustre-reviews/136343
      Kernel: 4.18.0-553.177.1.el8_10.x86_64

      Server
      Lustre: 2.17.61.87
      Branch: master
      Build: https://build.whamcloud.com/job/lustre-reviews/136343
      Kernel: 4.18.0-553.177.1.el8_10_lustre.x86_64

      sanity-quota test_84 (ZFS) fails intermittently with "dd succeed, expect EDQUOT": the final 200M direct write by the quota user succeeds after the group block hardlimit was set to 100M, so the limit was not enforced yet.

      Failed: https://testing.whamcloud.com/sub_tests/edfa7e35-2e5c-4e6a-8ba6-6be94858ea83 (review-zfs, el8.10, master d6be6e4efe66, https://review.whamcloud.com/70472). Same error on 2026-11-04 on change 70348 (review-dne-zfs-part-4, session 91f5d673-2231-45da-b3f7-1676f403ae05). Rate is low: 2 failures in about 6500 lustre-reviews runs in 30 days; master-next full-part-2 also failed test_84 3 times in the last 10 days with other messages (dd failed, expect succeed; set user quota failed; setup quota failed).

      The OST dmesg of the failing run, just before the failure, shows:
      qqi_reint_delayed()) lustre-OST0001: Delaying reintegration for qtype:0 until pending updates are flushed.
      qsd_req_completion()) DQACQ failed with -22 ... qtype:grp id:63 enforced:0 ... default:yes
      So the group lqe on the OST was still enforced:0 when dd ran. The test sets "-B 100M" right after clearing and resetting quota (OBD_FAIL_QUOTA_GRANT, setquota -r) and runs dd with no wait_quota_synced on the new limit.

      This is not LU-18646 (Resolved, fix 57871 in the tested tree: ZFS dd not consuming space, fixed by using /dev/urandom); the symptom string is the same but the cause here is quota reintegration/sync timing. The patch under test (LU-21060 jobstats split) does not touch quota.

      Raised by Patch Watcher run pw-engineer-70472-ps3-8188f76fa8b3 on https://review.whamcloud.com/70472

      VVVVVVV DO NOT REMOVE LINES BELOW, Added by Maloo for auto-association VVVVVVV
      sanity-quota test_84 - dd succeed, expect EDQUOT

            wc-triage WC Triage
            maloo Maloo
            Votes:
            0 Vote for this issue
            Watchers:
            1 Start watching this issue

              Created:
              Updated: