Uploaded image for project: 'MariaDB Server'
  1. MariaDB Server
  2. MDEV-32909

Hang not caught by semaphore wait threshold

    XMLWordPrintable

Details

    • Bug
    • Status: Closed (View Workflow)
    • Critical
    • Resolution: Won't Fix
    • 10.11.6
    • N/A
    • Storage Engine - InnoDB
    • Debian 11 Bullseye; MariaDB from the Maria repositories.

    Description

      Hello,

      Last night right after midnight one of my servers seemed to hang. All (or at least a great many?) queries would hang forever in the Execute phase. Of course this hung up the entire Galera cluster as well.

      I was awakened by an alert around 12:40am. My innodb_fatal_semaphore_wait_threshold is set to 32 seconds, so this hang was not caught by that watchdog.

      At 12:49 I was able to send MariaDB a signal which caused it to crash and dump core. So I do have a stack trace for this situation which I don't want to post publicly, but which is available upon request. I also have SHOW PROCESSLIST logs for much of the time in case that helps.

      If a MariaDB expert could take a look at the situation I would appreciate it! Thanks.

      Attachments

        Issue Links

          duplicates

          Bug - A problem which impairs or prevents the functions of the product. MDEV-32371 Deadlock between buf_page_get_zip() and buf_pool_t::corrupted_evict() on InnoDB ROW_FORMAT=COMPRESSED table corruption

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

          Activity

            People

              marko Marko Mäkelä
              xan@biblionix.com Xan Charbonnet
              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.