Uploaded image for project: 'MariaDB Server'
  1. MariaDB Server
  2. MDEV-35670

Table indexes suddenly corrupted after upgrading to 10.11.3

    XMLWordPrintable

Details

    Description

      We upgraded from 10.4.15 to 10.11.3 on November 09.

      After the upgrade, the index of the Master node table was corrupted twice.
      Both times, it was the same table, and after dropping the table, we dumped the data of the Slave and recovered it.

      2024-11-20  2:04:11 4299263 [ERROR] mariadbd: Index for table 'message' is corrupt; try to repair it
      2024-11-20  2:04:11 4299257 [ERROR] InnoDB: We detected index corruption in an InnoDB type table. You have to dump + drop + reimport the table or, in a case of widespread corruption, dump all InnoDB tables and recreate the whole tablespace. If the mariadbd server crashes after the startup or when you dump the tables. Please refer to https://mariadb.com/kb/en/library/innodb-recovery-modes/ for information about forcing recovery.
      

      2024-12-07  1:57:45 11642302 [ERROR] mariadbd: Index for table 'message' is corrupt; try to repair it
      2024-12-07  1:57:45 11642301 [ERROR] InnoDB: We detected index corruption in an InnoDB type table. You have to dump + drop + reimport the table or, in a case of widespread corruption, dump all InnoDB tables and recreate the whole tablespace. If the mariadbd server crashes after the startup or when you dump the tables. Please refer to https://mariadb.com/kb/en/library/innodb-recovery-modes/ for information about forcing recovery.
      

      The table is about 100G and is used only for batches.
      The batch query is performed in 0.1 seconds at 02:00 and is performed a total of 5 times.

      In this regard, I attach my.cnf, table syntax, and batch query.

      Since it is an operational environment, it is impossible to upgrade to 10.11.10 without identifying the exact cause.

      • Can searching and updating with multiple (320) PKs in the IN clause cause problems?
      • Can it be affected by the column order of the message_published_idx index (PK column is last)?
      • Is there a way to avoid index corrupt by changing settings, etc.?
      • Is there a way to reproduce the index corruption?

      Attachments

        1. my.cnf.TXT
          6 kB
          baek seung ho
        2. query.txt
          0.8 kB
          baek seung ho
        3. table.txt
          0.7 kB
          baek seung ho

        Activity

          People

            Unassigned Unassigned
            supbaek baek seung ho
            Votes:
            0 Vote for this issue
            Watchers:
            6 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.