Details

    • Bug
    • Status: Open (View Workflow)
    • Critical
    • Resolution: Unresolved
    • 10.11.9
    • 10.11
    • Galera
    • None

    Description

      Last night we encountered a MDL BF-BF conflict on the second node of one of our clusters.

      logging for the error:
      Feb 18 21:50:08 node02 mariadbd[1071]: 2025-02-18 21:50:08 9 [Note] WSREP: MDL BF-BF conflict
      Feb 18 21:50:08 node02 mariadbd[1071]: schema: db
      Feb 18 21:50:08 node02 mariadbd[1071]: request: (9 #011seqno 113068570 #011wsrep (high priority, exec, executing) cmd 0 161 #011update `c` set `s` = 'de4def60-4e1a-4c44-b2cc-544a02119308', `c`.`updated_at` = '2025-02-18T20:50:07Z' where `id` = '889d45d4-1617-4d72-83ff-28fc241b156b'^?ò´g#023#001)
      Feb 18 21:50:08 node02 mariadbd[1071]: granted: (8 #011seqno 113068568 #011wsrep (toi, exec, committed) cmd 0 45 #011OPTIMIZE TABLE `db`.`c_d_t`)
      Feb 18 21:50:08 node02 mariadbd[1071]: 2025-02-18 21:50:08 9 [ERROR] Aborting

      after that the node set wsrep_ready to "off"

      I've spoken to the developers, the tables in the 2 queries are not related (no triggers or anything affecting the table that is being optimized). From what I've gathered from the developers the OPTIMIZE is part of cronjob for multiple tables and multiple dbs. This cron started at 21:48:02
      The OPTIMIZE started at the first node at 21:50:07 and finished in 1.2024 seconds.

      unfortunately logging does not show any additional information.

      Attachments

        Issue Links

          Activity

            rtuk Rick Tuk added a comment - - edited

            2 days ago I've update our clusters from 10.11.9 to 10.11.11, last night we ran into the MDL BF-BF conflict again on 10.11.11

            logs from last night:

            Mar  4 21:51:31 node02 mariadbd[1063]: 2025-03-04 21:51:31 2 [Note] WSREP: MDL BF-BF conflict
            Mar  4 21:51:31 node02 mariadbd[1063]: schema:  db
            Mar  4 21:51:31 node02 mariadbd[1063]: request: (2 #011seqno 107959336 #011wsrep (high priority, exec, executing) cmd 0 161 #011insert into `c` (`s`, `st`, `cf`, `n`, `d`, `dd`, `bid`, `id`, `updated_at`, `created_at`) values ('p', 'custom', '[{\"id\":\"eaea1c10-a028-40f3-9e44-6cef1fafa02e\",\"name\":\"Projectnr\",\"type\":\"number\",\"priority\":7,\"value\":250220},{\"id\":\"72b62627-f963-47cb-9192-cbc0142805a4\",\"name\":\"Fasenr\",\"type\":\"number\",\"priority\":8,\"value\":38131},{\"id\":\"c7df13c5-bca
            Mar  4 21:51:31 node02 mariadbd[1063]: 2025-03-04 21:51:31 2 [ERROR] Aborting
            

            rtuk Rick Tuk added a comment - - edited 2 days ago I've update our clusters from 10.11.9 to 10.11.11, last night we ran into the MDL BF-BF conflict again on 10.11.11 logs from last night: Mar 4 21:51:31 node02 mariadbd[1063]: 2025-03-04 21:51:31 2 [Note] WSREP: MDL BF-BF conflict Mar 4 21:51:31 node02 mariadbd[1063]: schema: db Mar 4 21:51:31 node02 mariadbd[1063]: request: (2 #011seqno 107959336 #011wsrep (high priority, exec, executing) cmd 0 161 #011insert into `c` (`s`, `st`, `cf`, `n`, `d`, `dd`, `bid`, `id`, `updated_at`, `created_at`) values ('p', 'custom', '[{\"id\":\"eaea1c10-a028-40f3-9e44-6cef1fafa02e\",\"name\":\"Projectnr\",\"type\":\"number\",\"priority\":7,\"value\":250220},{\"id\":\"72b62627-f963-47cb-9192-cbc0142805a4\",\"name\":\"Fasenr\",\"type\":\"number\",\"priority\":8,\"value\":38131},{\"id\":\"c7df13c5-bca Mar 4 21:51:31 node02 mariadbd[1063]: 2025-03-04 21:51:31 2 [ERROR] Aborting
            janlindstrom Jan Lindström added a comment -

            rtuk Can you provide show create table for tables db.c and db.c_d_t. error logs and node configuration.

            janlindstrom Jan Lindström added a comment - rtuk Can you provide show create table for tables db.c and db.c_d_t. error logs and node configuration.
            rtuk Rick Tuk added a comment -

            I've added a zip with 2 directories, the show create tables for the tables and a directory with the config files from the cluster
            I do not have additional error logs, all information that was logged is already provided in this issue

            rtuk Rick Tuk added a comment - I've added a zip with 2 directories, the show create tables for the tables and a directory with the config files from the cluster I do not have additional error logs, all information that was logged is already provided in this issue
            janlindstrom Jan Lindström added a comment -

            rtuk Thank you. As expected there is foreign key references between these two tables and they cause MDL conflict. In this particular case OPTIMIZE should be allowed to continue and update should wait.

            janlindstrom Jan Lindström added a comment - rtuk Thank you. As expected there is foreign key references between these two tables and they cause MDL conflict. In this particular case OPTIMIZE should be allowed to continue and update should wait.

            People

              hemantdangi Hemant Dangi
              rtuk Rick Tuk
              Votes:
              0 Vote for this issue
              Watchers:
              5 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.