Uploaded image for project: 'MariaDB Server'
  1. MariaDB Server
  2. MDEV-25569

Restore crashed due to Data structure corruption

    XMLWordPrintable

Details

    Description

      Hello,

      Every day, I create a backup using MariaBackup, and restore the snapshot it on a Docker container.

      The version in both the main server and the docker container is 10.5.9.

      Usually, things go well, but recently I got two similar crashes due to "data structure corruption".
      I confirm the version was 10.5.9 in both sides, on both incidents.

      (many of these "Ignoring data file" logs)
      2021-04-19 7:33:01 0 [Note] InnoDB: Ignoring data file 'database1/table1.ibd' with space ID 20668, since the redo log references database1/table1.ibd with space ID 20666.
      2021-04-19 7:33:01 0 [Note] InnoDB: Ignoring data file './database1/table1_new.ibd' with space ID 20668. Another data file called database1/table1.ibd exists with the same space ID.
      2021-04-19 7:33:01 0 [Note] InnoDB: Starting final batch to recover 153185 pages from redo log.
      2021-04-19 7:33:01 0 [ERROR] InnoDB: Not applying INSERT_REUSE_REDUNDANT due to corruption on [page id: space=0, page number=7848]
      2021-04-19 7:33:01 0 [ERROR] InnoDB: Set innodb_force_recovery=1 to ignore corruption.
      2021-04-19 7:33:01 0 [ERROR] InnoDB: Not applying INSERT_HEAP_REDUNDANT due to corruption on [page id: space=0, page number=7848]
      2021-04-19 7:33:01 0 [ERROR] InnoDB: Set innodb_force_recovery=1 to ignore corruption.
      2021-04-19 7:33:08 0 [Note] InnoDB: To recover: 73635 pages from log
      2021-04-19 7:33:20 0 [ERROR] InnoDB: Plugin initialization aborted with error Data structure corruption
      [00] FATAL ERROR: 2021-04-19 07:33:20 mariabackup: innodb_init() returned 39 (Data structure corruption).

      –

      (many of these "Ignoring data file" logs)
      2021-04-29 7:32:31 0 [Note] InnoDB: Ignoring data file 'database1/table1.ibd' with space ID 26720, since the redo log references database1/table1.ibd with space ID 26718.
      2021-04-29 7:32:31 0 [Note] InnoDB: Ignoring data file 'database1/table2.ibd' with space ID 26721, since the redo log references database1/table2.ibd with space ID 26719.
      2021-04-29 7:32:31 0 [Note] InnoDB: Starting final batch to recover 116118 pages from redo log.
      2021-04-29 7:32:32 0 [ERROR] InnoDB: Not applying INSERT_REUSE_REDUNDANT due to corruption on [page id: space=0, page number=14132]
      2021-04-29 7:32:32 0 [ERROR] InnoDB: Set innodb_force_recovery=1 to ignore corruption.
      2021-04-29 7:32:43 0 [Note] InnoDB: To recover: 28397 pages from log
      2021-04-29 7:32:45 0 [ERROR] InnoDB: Plugin initialization aborted with error Data structure corruption
      [00] FATAL ERROR: 2021-04-29 07:32:45 mariabackup: innodb_init() returned 39 (Data structure corruption).

      Is there any way I can identify the problem here, or at least prevent this from happening?
      I want to ensure all my backups are "clean" and "valid".

      This may be related to the issue I reported about using "RENAME TABLE", MDEV-25568
      It was in April 15 that I added a job that performs "RENAME TABLE" every 5 minutes on two tables, so I wonder if the fact that this incident happened twice after that day, may be the cause of this.

      Thank you.

      Attachments

        Issue Links

          Activity

            People

              Unassigned Unassigned
              nunop Nuno
              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.