Uploaded image for project: 'MariaDB Server'
  1. MariaDB Server
  2. MDEV-16946

innodb.alter_kill failed in buildbot with wrong result

Details

    Description

      http://buildbot.askmonty.org/buildbot/builders/win32-packages/builds/9251

      innodb.alter_kill 'innodb'               w1 [ fail ]
      		 
              Test ended at 2018-08-03 06:41:41
      		 
       
      		 
      CURRENT_TEST: innodb.alter_kill
      		 
      --- D:/win32-packages/build/src/mysql-test/suite/innodb/r/alter_kill.result	2018-08-03 05:45:07.000000000 +0000
      		 
      +++ D:\win32-packages\build\src\mysql-test\suite\innodb\r\alter_kill.reject	2018-08-03 06:41:25.977718200 +0000
      		 
      @@ -62,7 +62,7 @@
      		 
       # Kill the server
      		 
       disconnect con1;
      		 
       # Attempt to start without an *.ibd file.
      		 
      -FOUND 1 /\[ERROR\] InnoDB: Tablespace [0-9]+ was not found at .*test.bug16735660.ibd/ in mysqld.1.err
      		 
      +NOT FOUND /\[ERROR\] InnoDB: Tablespace [0-9]+ was not found at .*test.bug16735660.ibd/ in mysqld.1.err
      		 
       SET TRANSACTION ISOLATION LEVEL READ UNCOMMITTED;
      		 
       SELECT * FROM bug16735660;
      		 
       a
      		 
       
      		 
      mysqltest: Result length mismatch

      Attachments

        Activity

          Ascending order - Click to sort in descending order
          elenst Elena Stepanova created issue -
          marko Marko Mäkelä added a comment -

          The failure was caused by a redo log checkpoint that occurred at a critical moment before the kill-and-restart. I was able to reproduce the failure by adding

          SET GLOBAL innodb_log_checkpoint_now=ON;

          immediately before the kill. (The parameter is only available in debug builds.)

          I fixed it so that the test will be skipped if a log checkpoint occurs at the critical moment.

          marko Marko Mäkelä added a comment - The failure was caused by a redo log checkpoint that occurred at a critical moment before the kill-and-restart. I was able to reproduce the failure by adding SET GLOBAL innodb_log_checkpoint_now= ON ; immediately before the kill. (The parameter is only available in debug builds.) I fixed it so that the test will be skipped if a log checkpoint occurs at the critical moment.
          marko Marko Mäkelä made changes -
          Field Original Value New Value
          issue.field.resolutiondate 2018-10-10 03:35:13.0 2018-10-10 03:35:13.267
          marko Marko Mäkelä made changes -
          Fix Version/s 10.3.11 [ 23141 ]
          Fix Version/s 10.2.19 [ 23207 ]
          Fix Version/s 10.2 [ 14601 ]
          Resolution Fixed [ 1 ]
          Status Open [ 1 ] Closed [ 6 ]
          serg Sergei Golubchik made changes -
          Workflow MariaDB v3 [ 88849 ] MariaDB v4 [ 154777 ]

          People

            marko Marko Mäkelä
            elenst Elena Stepanova
            Votes:
            0 Vote for this issue
            Watchers:
            2 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.