EVOLUTION-MANAGER

Edit File: multisource.inc

#
# This include file is used by more than one test suite
# (currently multisource and binlog_encryption).
# Please check all dependent tests after modifying it
#
# Usage:
# --source suite/rpl/include/multisource.inc
#
# By default, the script expects the length of the 2nd binary log to be
# $binlog_start_pos + length(Gtid_list event) + 2 x length(Binlog_checkpoint event)
# Some tests can have specific configuration which would change it,

#
# Test basic replication functionality 
# in multi-source setup
#

--source include/not_embedded.inc
--source include/have_innodb.inc
--source include/binlog_start_pos.inc
--let $rpl_server_count= 0

--connect (slave,127.0.0.1,root,,,$SERVER_MYPORT_3)

# MDEV-3984: crash/read of freed memory when changing master with named connection
# This fails after adding the new master 'abc', check we do not free twice.
--error ER_RELAY_LOG_INIT
change master 'abc' to relay_log_file='';
# This fails before adding the new master, check that we do free it.
--error ER_WRONG_ARGUMENTS
change master 'abc2' to master_host='';

# Start replication from the first master

--replace_result $SERVER_MYPORT_1 MYPORT_1
eval change master 'master1' to 
master_port=$SERVER_MYPORT_1, 
master_host='127.0.0.1', 
master_user='root';

start slave 'master1';
set default_master_connection = 'master1';
--source include/wait_for_slave_to_start.inc

--connect (master1,127.0.0.1,root,,,$SERVER_MYPORT_1)
--save_master_pos

--connection slave
--sync_with_master 0,'master1'

# Here and further: add an extra check on SQL thread status
# as the normal sync is not always enough
--source include/wait_for_sql_thread_read_all.inc

# each of the 3 commands should produce 
# 'master1' status

let $wait_for_all= 1; 
let $show_statement= SHOW ALL SLAVES STATUS;
let $field= Slave_IO_State;
let $condition= = 'Waiting for master to send event';
--source include/wait_show_condition.inc

--echo #
--echo # Checking SHOW SLAVE 'master1' STATUS
--echo #
--let $status_items= Master_Port, Relay_Log_File, Slave_IO_Running, Slave_SQL_Running, Last_Errno, Last_SQL_Errno
--let $slave_field_result_replace= /$SERVER_MYPORT_1/MYPORT_1/
--let $slave_name= 'master1'
--source include/show_slave_status.inc
--let $slave_name=

--echo #
--echo # Checking SHOW SLAVE STATUS
--echo #
--source include/show_slave_status.inc

--echo #
--echo # Checking SHOW ALL SLAVES STATUS
--echo #
--let $all_slaves_status= 1
--let $status_items= Connection_name, Master_Port, Relay_Log_File, Slave_IO_Running, Slave_SQL_Running, Last_Errno, Last_SQL_Errno, Slave_heartbeat_period
--source include/show_slave_status.inc
--let $all_slaves_status=
--echo #

# Check that replication actually works

--connection master1

--disable_warnings
drop database if exists db1;
--enable_warnings
create database db1;
use db1;
create table t1 (i int auto_increment, f1 varchar(16), primary key pk (i,f1)) engine=MyISAM;
insert into t1 (f1) values ('one'),('two');
--save_master_pos

--connection slave
--sync_with_master 0,'master1'

--sorted_result
select * from db1.t1;

--let $datadir = `SELECT @@datadir`

--echo # List of relay log files in the datadir
--list_files $datadir mysqld-relay-bin-master1.*

# Check that relay logs are recognizable

let binlog_start=4;
let binlog_file=;
source include/show_relaylog_events.inc;
let binlog_file= mysqld-relay-bin-master1.000002;
source include/show_relaylog_events.inc;

# Try to configure connection with the same name again,
# should get an error because the slave is running

--replace_result $SERVER_MYPORT_2 MYPORT_2
--error ER_SLAVE_MUST_STOP
eval change master 'master1' to
master_port=$SERVER_MYPORT_2,
master_host='127.0.0.1',
master_user='root';

# Try to configure using the default connection name
# (which is 'master1' at the moment),
# again, should get an error

--replace_result $SERVER_MYPORT_2 MYPORT_2
--error ER_SLAVE_MUST_STOP
eval change master to
master_port=$SERVER_MYPORT_2,
master_host='127.0.0.1',
master_user='root';

# Try to configure a connection with the same master
# using a different name, should get a conflict
 
--replace_result $SERVER_MYPORT_1 MYPORT_1
--error ER_CONNECTION_ALREADY_EXISTS
eval change master 'master2' to
master_port=$SERVER_MYPORT_1,
master_host='127.0.0.1',
master_user='root';

# Set up a proper 'default' connection to master2

set default_master_connection = '';

--replace_result $SERVER_MYPORT_2 MYPORT_2
eval change master to
master_port=$SERVER_MYPORT_2,
master_host='127.0.0.1',
master_user='root';

start slave;
--source include/wait_for_slave_to_start.inc

--source include/wait_for_sql_thread_read_all.inc

# See both connections in the same status output

let $wait_for_all= 1; 
let $show_statement= SHOW ALL SLAVES STATUS;
let $field= Slave_IO_State;
let $condition= = 'Waiting for master to send event';
--source include/wait_show_condition.inc

--echo #
--echo # Checking SHOW ALL SLAVES STATUS
--echo #
--let $all_slaves_status= 1
--let $status_items= Connection_name, Master_Port, Relay_Log_File, Slave_IO_Running, Slave_SQL_Running, Last_Errno, Last_SQL_Errno, Slave_heartbeat_period
--let $slave_field_result_replace= /$SERVER_MYPORT_1/MYPORT_1/ /$SERVER_MYPORT_2/MYPORT_2/
--source include/show_slave_status.inc
--let $all_slaves_status=
--echo #

# Check that replication from two servers actually works

--connection master1

insert into t1 (f1) values ('three');
--save_master_pos

--connect (master2,127.0.0.1,root,,,$SERVER_MYPORT_2)

--disable_warnings
drop database if exists db2;
--enable_warnings
create database db2;
use db2;
create table t1 (pk int auto_increment primary key, f1 int) engine=InnoDB;
begin;
insert into t1 (f1) values (1),(2);

--connection slave
--sync_with_master 0,'master1'

--connection master2
--save_master_pos

--connection slave
--sync_with_master 0
--sorted_result
select * from db1.t1;
select * from db2.t1;

--connection master2
commit;
--save_master_pos

--connection slave
--sync_with_master 0
--sorted_result
select * from db2.t1;

# Flush and purge logs on one master,
# make sure slaves don't get confused

--connection master1
flush logs;
--source include/wait_for_binlog_checkpoint.inc
--save_master_pos
--connection slave
--sync_with_master 0, 'master1'

--connection master1
purge binary logs to 'master-bin.000002';
# Additional events: 43 (Gtid_list) + 2 x 44 (Binlog_checkpoint) = 131
let filesize=`select $binlog_start_pos+131`;
--replace_result $filesize filesize
show binary logs;
insert into t1 (f1) values ('four');
create table db1.t3 (f1 int) engine=InnoDB;
--save_master_pos

--connection slave
--sync_with_master 0,'master1'

--source include/wait_for_sql_thread_read_all.inc

let $wait_for_all= 1; 
let $show_statement= SHOW ALL SLAVES STATUS;
let $field= Slave_IO_State;
let $condition= = 'Waiting for master to send event';
--source include/wait_show_condition.inc

--echo #
--echo # Checking SHOW ALL SLAVES STATUS
--echo #
--let $all_slaves_status= 1
--let $status_items= Connection_name, Master_Port, Relay_Log_File, Slave_IO_Running, Slave_SQL_Running, Last_Errno, Last_SQL_Errno, Slave_heartbeat_period
--let $slave_field_result_replace= /$SERVER_MYPORT_1/MYPORT_1/ /$SERVER_MYPORT_2/MYPORT_2/
--source include/show_slave_status.inc
--let $all_slaves_status=
--echo #

--sorted_result
select * from db1.t1;

# This should show relay log events for the default master
# (the one with the empty name)
let binlog_file=;
source include/show_relaylog_events.inc;
let binlog_file= mysqld-relay-bin.000002;
source include/show_relaylog_events.inc;

# Make sure we don't lose control over replication connections
# after reconnecting to the slave

--disconnect slave
--connect (slave,127.0.0.1,root,,,$SERVER_MYPORT_3)

stop slave io_thread;
show status like 'Slave_running';
set default_master_connection = 'master1';
show status like 'Slave_running';

# Cleanup

drop database db1;
drop database db2;

--source include/reset_master_slave.inc
--disconnect slave

--connection master1
drop database db1;
--source include/reset_master_slave.inc
--disconnect master1

--connection master2
drop database db2;
--source include/reset_master_slave.inc
--disconnect master2