Uploaded image for project: 'MariaDB Server'
  1. MariaDB Server
  2. MDEV-4474

Replication aborts on starting a circular replication using GTID

    XMLWordPrintable

Details

    • Bug
    • Status: Closed (View Workflow)
    • Minor
    • Resolution: Fixed
    • 10.0.2
    • 10.0.3
    • None
    • None

    Description

      I set replication 1->2 to use GTID, start it, execute some events on server 1 and server 2, then set replication 2->1 to use GTID too, and attempt to start it.
      In the example below, it fails with "'Table 'test.t2' doesn't exist'", apparently it misses an event upon startup, although it's present in the binary log.

      Output of the test case provided below

      # 
      # For now we'll only have 1->2 running
      # 
      # Server 1
      # Stop replication 2->1
      include/stop_slave.inc
      # 
      # Server 2
      # Use GTID for replication 1->2
      include/stop_slave.inc
      change master to master_use_gtid=1;
      include/start_slave.inc
      # 
      # Create some 0-1-* and 0-2-* events in binlog of server 2
      connection server_1;
      create table t1 (i int) engine=InnoDB;
      insert into t1 values (1);
      connection server_2;
      create table t2 (i int) engine=InnoDB;
      connection server_1;
      insert into t1 values (2);
      connection server_2;
      insert into t2 values (1);
      # 
      # All events are present in the binlog of server 2
      show binlog events;
      Log_name	Pos	Event_type	Server_id	End_log_pos	Info
      slave-bin.000001	4	Format_desc	2	248	Server ver: 10.0.1-MariaDB-debug-log, Binlog ver: 4
      slave-bin.000001	248	Gtid_list	2	271	[]
      slave-bin.000001	271	Binlog_checkpoint	2	310	slave-bin.000001
      slave-bin.000001	310	Gtid	1	348	GTID 0-1-1
      slave-bin.000001	348	Query	1	453	use `test`; create table t1 (i int) engine=InnoDB
      slave-bin.000001	453	Gtid	1	491	BEGIN GTID 0-1-2
      slave-bin.000001	491	Query	1	584	use `test`; insert into t1 values (1)
      slave-bin.000001	584	Xid	1	611	COMMIT /* xid=277 */
      slave-bin.000001	611	Gtid	2	649	GTID 0-2-3
      slave-bin.000001	649	Query	2	754	use `test`; create table t2 (i int) engine=InnoDB
      slave-bin.000001	754	Gtid	1	792	BEGIN GTID 0-1-3
      slave-bin.000001	792	Query	1	885	use `test`; insert into t1 values (2)
      slave-bin.000001	885	Xid	1	912	COMMIT /* xid=282 */
      slave-bin.000001	912	Gtid	2	950	BEGIN GTID 0-2-4
      slave-bin.000001	950	Query	2	1043	use `test`; insert into t2 values (1)
      slave-bin.000001	1043	Xid	2	1070	COMMIT /* xid=283 */
      # 
      # Server 1
      # Start replication 2->1 using GTID, 
      # it fails with 'Table 'test.t2' doesn't exist'
      # (which shows up either as a failure on sync_with_master,
      #  or more often as hanging start_slave.inc)
      change master to master_use_gtid=1;
      include/start_slave.inc

      MariaDB [test]> show slave status \G
      *************************** 1. row ***************************
                     Slave_IO_State: Waiting for master to send event
                        Master_Host: 127.0.0.1
                        Master_User: root
                        Master_Port: 16001
                      Connect_Retry: 1
                    Master_Log_File: slave-bin.000001
                Read_Master_Log_Pos: 1070
                     Relay_Log_File: master-relay-bin.000002
                      Relay_Log_Pos: 597
              Relay_Master_Log_File: slave-bin.000001
                   Slave_IO_Running: Yes
                  Slave_SQL_Running: No
                    Replicate_Do_DB: 
                Replicate_Ignore_DB: 
                 Replicate_Do_Table: 
             Replicate_Ignore_Table: 
            Replicate_Wild_Do_Table: 
        Replicate_Wild_Ignore_Table: 
                         Last_Errno: 1146
                         Last_Error: Error 'Table 'test.t2' doesn't exist' on query. Default database: 'test'. Query: 'insert into t2 values (1)'
                       Skip_Counter: 0
                Exec_Master_Log_Pos: 310
                    Relay_Log_Space: 1053
                    Until_Condition: None
                     Until_Log_File: 
                      Until_Log_Pos: 0
                 Master_SSL_Allowed: No
                 Master_SSL_CA_File: 
                 Master_SSL_CA_Path: 
                    Master_SSL_Cert: 
                  Master_SSL_Cipher: 
                     Master_SSL_Key: 
              Seconds_Behind_Master: NULL
      Master_SSL_Verify_Server_Cert: No
                      Last_IO_Errno: 0
                      Last_IO_Error: 
                     Last_SQL_Errno: 1146
                     Last_SQL_Error: Error 'Table 'test.t2' doesn't exist' on query. Default database: 'test'. Query: 'insert into t2 values (1)'
        Replicate_Ignore_Server_Ids: 
                   Master_Server_Id: 2
                         Using_Gtid: 1
      1 row in set (0.00 sec)

      Test case:

      --source include/have_innodb.inc
      --let $rpl_topology=1->2->1
      --source include/rpl_init.inc
       
      --echo # 
      --echo # For now we'll only have 1->2 running
       
      --echo # 
      --echo # Server 1
      --echo # Stop replication 2->1
      --connection server_1
      --source include/stop_slave.inc
       
      --echo # 
      --echo # Server 2
      --echo # Use GTID for replication 1->2
      --connection server_2
      --source include/stop_slave.inc
      change master to master_use_gtid=1;
      --source include/start_slave.inc
       
      --echo # 
      --echo # Create some 0-1-* and 0-2-* events in binlog of server 2
       
      --enable_connect_log
       
      --connection server_1
      create table t1 (i int) engine=InnoDB;
      insert into t1 values (1);
      --save_master_pos
       
      --connection server_2
      --sync_with_master
      create table t2 (i int) engine=InnoDB;
      --save_master_pos
       
      --connection server_1
      insert into t1 values (2);
      --save_master_pos
       
      --connection server_2
      --sync_with_master
      insert into t2 values (1);
      --save_master_pos
       
      --disable_connect_log
       
      --echo # 
      --echo # All events are present in the binlog of server 2
       
      show binlog events;
       
      --echo # 
      --echo # Server 1
      --echo # Start replication 2->1 using GTID, 
      --echo # it fails with 'Table 'test.t2' doesn't exist'
      --echo # (which shows up either as a failure on sync_with_master,
      --echo #  or more often as hanging start_slave.inc)
       
      --connection server_1
      change master to master_use_gtid=1;
      --source include/start_slave.inc
      --sync_with_master
       
       
      --source include/rpl_end.inc

      cnf file:

      !include suite/rpl/rpl_1slave_base.cnf
      !include include/default_client.cnf
       
       
      [mysqld.1]
      log-slave-updates
      loose-innodb
       
      [mysqld.2]
      log-slave-updates
      loose-innodb

      bzr version-info

      revision-id: knielsen@knielsen-hq.org-20130503092729-gedp152b19k5wdnj
      revno: 3626
      branch-nick: 10.0-base

      Attachments

        Issue Links

          Activity

            People

              knielsen Kristian Nielsen
              elenst Elena Stepanova
              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.