Uploaded image for project: 'MariaDB Server'
  1. MariaDB Server
  2. MDEV-7998

Total lockup on two add index statements

    XMLWordPrintable

Details

    • Bug
    • Status: Closed (View Workflow)
    • Major
    • Resolution: Won't Fix
    • 10.0.17-galera
    • N/A
    • Galera
    • CentOS 5.11

    Description

      We have galera with two nodes - main one with all access and second as active backup and reports server.
      I was running alter table add index on our main node (did it already on the other node) with
      SET GLOBAL wsrep_OSU_method='RSU';
      SET SESSION wsrep_on=OFF;

      I was monitoring progress in other window and instead of "show processlist;" I unfortunately ran from history older alter table add index on different table - it should have reported error on duplicating index, but instead entire node locked, no connection (new or already established) was responding. I had to ask our server admin to restart it and he had to kill 9 it, then it started OK. Even the added index is present, so it seems to have finished in the background (there was enough time for it before stop was issued).

      Attached is relevant part of mysqld.log, (few rows before event are from day before, I added couple of comments prefixed with !!!!

      Attachments

        1. mariadb_alter_errorlog.txt
          14 kB
        2. server.cnf
          2 kB
        3. tables.sql
          6 kB

        Activity

          People

            seppo Seppo Jaakola
            jkavalik Jiri Kavalik
            Votes:
            0 Vote for this issue
            Watchers:
            3 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.