Uploaded image for project: 'MariaDB Server'
  1. MariaDB Server
  2. MDEV-750

LP:800035 - intermittent rpl_deadlock_innodb failures

Details

    Description

      For example:

      http://jenkins.percona.com/job/percona-server-5.5-trunk/BUILD_TYPE=release,Host=ubuntu-maverick-64bit/18/console

      rpl.rpl_deadlock_innodb 'mix'            w7 [ retry-fail ]
              Test ended at 2011-06-21 03:34:19
       
      CURRENT_TEST: rpl.rpl_deadlock_innodb
      mysqltest: In included file "./include/wait_for_slave_param.inc": 
      included from ./include/wait_for_slave_param.inc at line 115:
      included from ./include/wait_for_slave_sql_error.inc at line 41:
      included from ./extra/rpl_tests/rpl_deadlock.test at line 84:
      At line 115: Timeout in include/wait_for_slave_param.inc
       
      The result from queries just before the failure was:
      < snip >
      master-bin.000001	845
       
      **** SHOW BINLOG EVENTS on server_1 ****
      binlog_name = 'master-bin.000001'
      SHOW BINLOG EVENTS IN 'master-bin.000001';
      Log_name	Pos	Event_type	Server_id	End_log_pos	Info
      master-bin.000001	4	Format_desc	1	107	Server ver: 5.5.13-log, Binlog ver: 4
      master-bin.000001	107	Query	1	224	use `test`; CREATE TABLE t1 (a INT NOT NULL, KEY(a)) ENGINE=innodb
      master-bin.000001	224	Query	1	324	use `test`; CREATE TABLE t2 (a INT) ENGINE=innodb
      master-bin.000001	324	Query	1	441	use `test`; CREATE TABLE t3 (a INT NOT NULL, KEY(a)) ENGINE=innodb
      master-bin.000001	441	Query	1	509	BEGIN
      master-bin.000001	509	Query	1	597	use `test`; INSERT INTO t1 VALUES (1)
      master-bin.000001	597	Query	1	730	use `test`; INSERT INTO t2 VALUES (2), (2), (2), (2), (2), (2), (2), (2), (2), (2)
      master-bin.000001	730	Query	1	818	use `test`; INSERT INTO t3 VALUES (3)
      master-bin.000001	818	Xid	1	845	COMMIT /* xid=384 */
       
      **** SHOW RELAYLOG EVENTS on server_1 ****
      relaylog_name = 'No such row'
      SHOW RELAYLOG EVENTS IN 'No such row';
      Log_name	Pos	Event_type	Server_id	End_log_pos	Info
       
      More results from queries before failure can be found in /mnt/workspace/percona-server-5.5-trunk/BUILD_TYPE/release/Host/ubuntu-maverick-64bit/Percona-Server-5.5.13-rel20.4/mysql-test/var/7/log/rpl_deadlock_innodb.log
       
       - saving '/mnt/workspace/percona-server-5.5-trunk/BUILD_TYPE/release/Host/ubuntu-maverick-64bit/Percona-Server-5.5.13-rel20.4/mysql-test/var/7/log/rpl.rpl_deadlock_innodb-mix/' to '/mnt/workspace/percona-server-5.5-trunk/BUILD_TYPE/release/Host/ubuntu-maverick-64bit/Percona-Server-5.5.13-rel20.4/mysql-test/var/log/rpl.rpl_deadlock_innodb-mix/'

      Another instance on a branch with unrelated change:
      http://jenkins.percona.com/job/percona-server-5.5-param/BUILD_TYPE=debug,Host=ubuntu-natty-64bit/8/console

      Attachments

        Activity

          stewart Stewart Smith added a comment - - edited

          Re: intermittent rpl_deadlock_innodb failures
          currently trying this patch to see if it may help:

          --- a/mysql-test/include/wait_for_slave_param.inc	2011-04-12 22:38:45.000000000 +1000
          +++ b/mysql-test/include/wait_for_slave_param.inc	2011-06-21 22:59:13.804854682 +1000
          @@ -79,7 +79,7 @@
           
           # mysqltest doesn't provide any better way to multiply by 10
           --let $_wait_for_slave_param_zero= 0
          ---let $_slave_timeout_counter= $_slave_timeout$zero
          +--let $_slave_timeout_counter= $_slave_timeout$_wait_for_slave_param_zero
           --let $_slave_continue= 1
           while ($_slave_continue)
           {

          stewart Stewart Smith added a comment - - edited Re: intermittent rpl_deadlock_innodb failures currently trying this patch to see if it may help: --- a/mysql-test/include/wait_for_slave_param.inc 2011-04-12 22:38:45.000000000 +1000 +++ b/mysql-test/include/wait_for_slave_param.inc 2011-06-21 22:59:13.804854682 +1000 @@ -79,7 +79,7 @@ # mysqltest doesn't provide any better way to multiply by 10 --let $_wait_for_slave_param_zero= 0 ---let $_slave_timeout_counter= $_slave_timeout$zero +--let $_slave_timeout_counter= $_slave_timeout$_wait_for_slave_param_zero --let $_slave_continue= 1 while ($_slave_continue) {
          stewart Stewart Smith added a comment -

          Re: intermittent rpl_deadlock_innodb failures
          Not sure if this affects Maria or not... but it's possible that you may be hitting it, depending on your build+test setup I guess.

          stewart Stewart Smith added a comment - Re: intermittent rpl_deadlock_innodb failures Not sure if this affects Maria or not... but it's possible that you may be hitting it, depending on your build+test setup I guess.

          Launchpad bug id: 800035

          ratzpo Rasmus Johansson (Inactive) added a comment - Launchpad bug id: 800035

          Still an issue?

          serg Sergei Golubchik added a comment - Still an issue?

          We had a failure like that about 6 weeks ago on 5.3 tree on opensolaris:
          http://buildbot.askmonty.org/buildbot/builders/opensolaris-511-bintar/builds/2896/steps/test/logs/stdio

          Before that, buildbot only returns similar failures for 2011, but as we know not all data gets stored.

          Looking at the include file, the obvious bug was fixed in 5.5 in the same way as suggested (as a part of revno 3178 / revision-id: knielsen@knielsen-hq.org-20111214110203-nl2k00i85g16cnvj), but it still exists in 5.1-5.3. I will apply it to the earlier trees as well, it's a one-line change in tests, why not do it if it helps to avoid a few sporadic failures.

          elenst Elena Stepanova added a comment - We had a failure like that about 6 weeks ago on 5.3 tree on opensolaris: http://buildbot.askmonty.org/buildbot/builders/opensolaris-511-bintar/builds/2896/steps/test/logs/stdio Before that, buildbot only returns similar failures for 2011, but as we know not all data gets stored. Looking at the include file, the obvious bug was fixed in 5.5 in the same way as suggested (as a part of revno 3178 / revision-id: knielsen@knielsen-hq.org-20111214110203-nl2k00i85g16cnvj), but it still exists in 5.1-5.3. I will apply it to the earlier trees as well, it's a one-line change in tests, why not do it if it helps to avoid a few sporadic failures.

          The fix for the typo pushed in 5.1 tree:
          http://bazaar.launchpad.net/~maria-captains/maria/5.1/revision/3220

          In 5.5 it was fixed long time ago (before 5.5.20) here:
          http://bazaar.launchpad.net/~maria-captains/maria/5.5/revision/3178

          elenst Elena Stepanova added a comment - The fix for the typo pushed in 5.1 tree: http://bazaar.launchpad.net/~maria-captains/maria/5.1/revision/3220 In 5.5 it was fixed long time ago (before 5.5.20) here: http://bazaar.launchpad.net/~maria-captains/maria/5.5/revision/3178

          People

            elenst Elena Stepanova
            laurynasbiveinis Laurynas Biveinis
            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.