Uploaded image for project: 'MariaDB Server'
  1. MariaDB Server
  2. MDEV-18590

galera.versioning_trx_id: Test failure: mysqltest: Result content mismatch

Details

    Description

      MariaDB Server: 10.4.3-MariaDB-debug built from sources, branch=10.4, commit: c568e25379600db8af4bd39df4761ba0fbc1a14e.
      Galera4 lib: debug built from sources, commit: 9cdbeb86c330b808571b14270e6428accb899c58.

      Run:

      ./mtr galera.versioning_trx_id --force --max-test-fail=0 --repeat=3
      ./mtr --suite galera --force --big-test --max-test-fail=0
      

      stdout.log: (the whole file is attached: 190215_desktop_stdout.log)

      Completed: Failed 1/3 tests, 66.67% were successful.
       
      CURRENT_TEST: galera.versioning_trx_id
      --- /home/stepan/galera/git/10.4/server/mysql-test/suite/galera/r/versioning_trx_id.result	2019-01-23 13:59:37.220821400 +0200
      +++ /home/stepan/galera/git/10.4/server/mysql-test/suite/galera/r/versioning_trx_id.reject	2019-02-15 15:38:57.358652400 +0200
      @@ -48,7 +48,7 @@
       0
       select count(*) from mysql.transaction_registry where begin_timestamp>=commit_timestamp;
       count(*)
      -0
      +1
       drop table t1;
       disconnect node_2;
       disconnect node_1;
       
      mysqltest: Result content mismatch
      

      See also attached 190215_galera.versioning_trx_id.zip.

      Note: This test is failed sporadically.

      Attachments

        Issue Links

          Activity

            stepan.patryshev Stepan Patryshev (Inactive) added a comment - It is still failed on BB: https://mdbe-ci-repo.mariadb.net/bb-logs/MariaDBServerCommunity/10.3-2020-Jun-21-10-00-00/rhel/8/mtr-galera-test/mtr.log

            kevg I have some questions on this test case:

            • What this test is trying to enforce ?
            • Have you tested this case using async replication?
            • When/Where exactly those begin_timestamp and commit_timestamp are computed ?
            • Writes to mysql.transaction_registry are not replicated by Galera
            • Above could mean that those timestamps on mysql.transaction_registry are computed on every node or not at all depending where the computation is
            • I'm not familiar with system versioned tables implementation
            jplindst Jan Lindström (Inactive) added a comment - kevg I have some questions on this test case: What this test is trying to enforce ? Have you tested this case using async replication? When/Where exactly those begin_timestamp and commit_timestamp are computed ? Writes to mysql.transaction_registry are not replicated by Galera Above could mean that those timestamps on mysql.transaction_registry are computed on every node or not at all depending where the computation is I'm not familiar with system versioned tables implementation

            Transactional versioning is low priority until new design.

            midenok Aleksey Midenkov added a comment - Transactional versioning is low priority until new design.
            janlindstrom Jan Lindström added a comment - https://github.com/MariaDB/server/pull/3192
            sysprg Julius Goryavsky added a comment - Fix merged with head revision: https://github.com/MariaDB/server/commit/136358036dfe2637b2db428b848148475ac21154

            People

              sysprg Julius Goryavsky
              stepan.patryshev Stepan Patryshev (Inactive)
              Votes:
              0 Vote for this issue
              Watchers:
              12 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved:

                Git Integration

                  Error rendering 'com.xiplink.jira.git.jira_git_plugin:git-issue-webpanel'. Please contact your Jira administrators.