mirror of
https://github.com/MariaDB/server.git
synced 2025-02-09 23:24:11 +01:00
eb4458e993
The current semi-sync binlog fail-over recovery process uses rpl_semi_sync_slave_enabled==TRUE as its condition to truncate a primary server’s binlog, as it is anticipating the server to re-join a replication topology as a replica. However, for servers configured with both rpl_semi_sync_master_enabled=1 and rpl_semi_sync_slave_enabled=1, if a primary is just re-started (i.e. retaining its role as master), it can truncate its binlog to drop transactions which its replica(s) has already received and executed. If this happens, when the replica reconnects, its gtid_slave_pos can be ahead of the recovered primary’s gtid_binlog_pos, resulting in an error state where the replica’s state is ahead of the primary’s. This patch changes the condition for semi-sync recovery to truncate the binlog to instead use the configuration variable --init-rpl-role, when set to SLAVE. This allows for both rpl_semi_sync_master_enabled and rpl_semi_sync_slave_enabled to be set for a primary that is restarted, and no transactions will be lost, so long as --init-rpl-role is not set to SLAVE. Reviewed By: ============ Sergei Golubchik <serg@mariadb.com>
124 lines
4.2 KiB
PHP
124 lines
4.2 KiB
PHP
if ($failover_to_slave)
|
|
{
|
|
--let $server_to_crash=1
|
|
--let $server_to_promote=2
|
|
--let $new_master_port=$SERVER_MYPORT_2
|
|
--let $client_port=$SERVER_MYPORT_1
|
|
|
|
--connect (conn_client,127.0.0.1,root,,test,$SERVER_MYPORT_1,)
|
|
}
|
|
if (!$failover_to_slave)
|
|
{
|
|
--let $server_to_crash=2
|
|
--let $server_to_promote=1
|
|
--let $new_master_port=$SERVER_MYPORT_1
|
|
--let $client_port=$SERVER_MYPORT_2
|
|
|
|
--connect (conn_client,127.0.0.1,root,,test,$SERVER_MYPORT_2,)
|
|
}
|
|
|
|
|
|
# Hold insert after write to binlog and before "run_commit_ordered" in engine
|
|
|
|
|
|
if ($case == 1)
|
|
{
|
|
SET DEBUG_SYNC= "commit_after_release_LOCK_after_binlog_sync SIGNAL con1_ready WAIT_FOR con1_go";
|
|
--send_eval $query_to_crash
|
|
--connection server_$server_to_crash
|
|
SET DEBUG_SYNC= "now WAIT_FOR con1_ready";
|
|
--source include/kill_mysqld.inc
|
|
}
|
|
|
|
# complicate recovery with an extra binlog file
|
|
if ($case == 2)
|
|
{
|
|
SET DEBUG_SYNC= "commit_before_get_LOCK_commit_ordered SIGNAL con1_ready WAIT_FOR con1_go";
|
|
SET DEBUG_SYNC= "commit_after_release_LOCK_after_binlog_sync WAIT_FOR con1_go1";
|
|
--send_eval $query_to_crash
|
|
--connect (conn_client_2,127.0.0.1,root,,test,$SERVER_MYPORT_2,)
|
|
# use the same signal with $query_to_crash
|
|
SET DEBUG_SYNC= "now WAIT_FOR con1_ready";
|
|
SET DEBUG_SYNC= "commit_before_get_LOCK_after_binlog_sync SIGNAL con1_go";
|
|
SET DEBUG_SYNC= "commit_before_get_LOCK_commit_ordered SIGNAL con2_ready";
|
|
--send_eval $query2_to_crash
|
|
--connection server_$server_to_crash
|
|
SET DEBUG_SYNC= "now WAIT_FOR con2_ready";
|
|
--source include/kill_mysqld.inc
|
|
}
|
|
|
|
# complicate recovery with an extra binlog file
|
|
if ($case == 3)
|
|
{
|
|
SET DEBUG_SYNC= "commit_before_get_LOCK_commit_ordered SIGNAL con1_ready WAIT_FOR con1_go";
|
|
--send_eval $query_to_crash
|
|
--connect (conn_client_3,127.0.0.1,root,,test,$SERVER_MYPORT_1,)
|
|
SET DEBUG_SYNC= "now WAIT_FOR con1_ready";
|
|
# use the same signal with $query_to_crash
|
|
SET DEBUG_SYNC= "commit_before_update_binlog_end_pos SIGNAL con3_ready WAIT_FOR con1_go";
|
|
--send_eval $query2_to_crash
|
|
--connection server_$server_to_crash
|
|
SET DEBUG_SYNC= "now WAIT_FOR con3_ready";
|
|
--source include/kill_mysqld.inc
|
|
}
|
|
|
|
--connection server_$server_to_promote
|
|
--let $slave_param= Slave_SQL_Running_State
|
|
--let $slave_param_value= Slave has read all relay log; waiting for more updates
|
|
source include/wait_for_slave_param.inc;
|
|
|
|
--error 2003
|
|
--source include/stop_slave.inc
|
|
|
|
--eval select count(*) 'on slave must be $expected_rows_on_slave' from t1
|
|
|
|
SELECT @@GLOBAL.gtid_current_pos;
|
|
|
|
--let $restart_parameters=--skip-slave-start=1 --rpl-semi-sync-slave-enabled=1 --init-rpl-role=SLAVE
|
|
--let $allow_rpl_inited=1
|
|
--source include/start_mysqld.inc
|
|
--connection server_$server_to_crash
|
|
--enable_reconnect
|
|
--source include/wait_until_connected_again.inc
|
|
|
|
--let $slave_semi_sync_enabled= query_get_value(SHOW VARIABLES LIKE 'Rpl_semi_sync_slave_enabled', Value, 1)
|
|
--echo # Ensuring variable rpl_semi_sync_slave_enabled is ON..
|
|
if (`SELECT strcmp("ON", "$slave_semi_sync_enabled") != 0`)
|
|
{
|
|
--die Slave started with rpl_semi_sync_slave_enabled=1 yet it is OFF in the variable output
|
|
}
|
|
|
|
--let $slave_semi_sync_status= query_get_value(SHOW STATUS LIKE 'Rpl_semi_sync_slave_status', Value, 1)
|
|
--echo # Ensuring status rpl_semi_sync_slave_status is OFF..
|
|
if (`SELECT strcmp("OFF", "$slave_semi_sync_status") != 0`)
|
|
{
|
|
--die Slave started with skip-slave-start yet started with rpl_semi_sync_slave_status=ON
|
|
}
|
|
|
|
--eval select count(*) 'on master must be $expected_rows_on_master' from t1
|
|
|
|
# Check error log for correct messages.
|
|
let $log_error_ = $MYSQLTEST_VARDIR/log/mysqld.$server_to_crash.err;
|
|
--let SEARCH_FILE=$log_error_
|
|
--let SEARCH_PATTERN=$log_search_pattern
|
|
--source include/search_pattern_in_file.inc
|
|
|
|
--disconnect conn_client
|
|
|
|
#
|
|
# FAIL OVER now to new master
|
|
#
|
|
--connection server_$server_to_promote
|
|
set global rpl_semi_sync_master_enabled = 1;
|
|
set global rpl_semi_sync_master_wait_point=AFTER_SYNC;
|
|
|
|
--connection server_$server_to_crash
|
|
--let $master_port=$SERVER_MYPORT_2
|
|
if (`select $server_to_crash = 2`)
|
|
{
|
|
--let $master_port=$SERVER_MYPORT_1
|
|
}
|
|
evalp CHANGE MASTER TO master_host='127.0.0.1', master_port=$new_master_port, master_user='root', master_use_gtid=SLAVE_POS;
|
|
set global rpl_semi_sync_slave_enabled = 1;
|
|
set @@global.gtid_slave_pos=@@global.gtid_binlog_pos;
|
|
--source include/start_slave.inc
|