Uploaded image for project: 'MariaDB Server'
  1. MariaDB Server
  2. MDEV-35226

InnoDB occasionally fails to recover a corrupted page from the doublewrite buffer

Details

    Description

      While testing the fix of MDEV-35225, I came up with the following rare failure of the test innodb.innodb-32k-crash:

      10.6 1067046b7f4337080e5b59c364c0a53cbeca715e

      		 
      2024-10-21 23:10:11 0 [Note] InnoDB: Completed initialization of buffer pool
      		 
      2024-10-21 23:10:11 0 [Note] InnoDB: Starting crash recovery from checkpoint LSN=1059287101,1062232449
      		 
      2024-10-21 23:10:11 0 [ERROR] InnoDB: Cannot apply log to [page id: space=704, page number=0] of corrupted file './test/t1.ibd'
      		 
      2024-10-21 23:10:11 0 [ERROR] InnoDB: Plugin initialization aborted at srv0start.cc[1513] with error Data structure corruption

      When I tried recovery on a copy of the data directory 32k-recovery.tar.xz, it would succeed each time. The difference to the above would be that there would be a message about doublewrite buffer recovery:

      10.6 1067046b7f4337080e5b59c364c0a53cbeca715e

      		 
      2024-10-21 23:42:51 0 [Note] InnoDB: Starting crash recovery from checkpoint LSN=1059287101,1059287101
      		 
      2024-10-21 23:42:51 0 [Note] InnoDB: Trying to recover page [page id: space=704, page number=0] from the doublewrite buffer.
      		 
      2024-10-21 23:42:51 0 [Note] InnoDB: Recovered page [page id: space=704, page number=0] to './test/t1.ibd' from the doublewrite buffer.
      		 
      2024-10-21 23:42:51 0 [Note] InnoDB: 1 transaction(s) which must be rolled back or cleaned up in total 20 row operations to undo
      		 
      2024-10-21 23:42:51 0 [Note] InnoDB: Trx id counter is 52417
      		 
      2024-10-21 23:42:51 0 [Note] InnoDB: To recover: 698 pages
      		 
      2024-10-21 23:42:51 0 [Note] InnoDB: Resizing redo log from 33554432 to 100663296 bytes; LSN=1062232813

      I will try to reproduce this failure under rr record so that this can be analyzed and fixed.

      Attachments

        Issue Links

          relates to

          Bug - A problem which impairs or prevents the functions of the product. MDEV-34830 LSN in the future is not being treated as serious corruption

          • Major - Major loss of function.
          • Closed

          Bug - A problem which impairs or prevents the functions of the product. MDEV-35225 Bogus debug assertion failures in innodb.innodb-32k-crash

          • Critical - Crashes, loss of data, severe memory leak.
          • Closed

          Activity

            Transition Time In Source Status Execution Times
            Marko Mäkelä made transition -
            Open In Progress
            		56s 1
            Marko Mäkelä made transition -
            In Progress Stalled
            		2h 9m 1

            People

              mleich Matthias Leich
              marko Marko Mäkelä
              Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

              Dates

                Created:
                Updated:

                Git Integration

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