mirror of
https://github.com/MariaDB/server.git
synced 2025-11-30 05:23:50 +03:00
475f69b985
Analysis:
========
In case multi binlog truncation scenario debug sync points are in the
following order.
Two inserts are done on master as shown below.
INSERT INTO t1 VALUES (4, REPEAT("x", 4100)
commit_after_release_LOCK_after_binlog_sync
INSERT INTO t1 VALUES (5, REPEAT("x", 4100)
commit_after_release_LOCK_log
First insert debug sync ensures that transaction is synced to binlog and
not committed but it reached slave through semi sync.
Second insert debug sync ensures that transaction is synced to binlog and
not committed. It doesn't ensure that 'INSERT 5' reached slave.
Most of the times INSERT 5 reaches slave, hence when it is promoted as
master it sends 4,5 to slave. But occasionally 5 may not reach slave in
those cases post recovery master will have only 4. When row 6 is inserted
Master has 4-6 and Slave has 4,5,6.
This results in test failure.
Fix:
===
For the first insert use 'commit_before_get_LOCK_commit_ordered' debug sync
point, it will ensure that binlog was sent to slave and slave has
acknowledged the receipt. Now enable debug code such that when the next
transaction is written to binary log, dump thread will read and send it
across the network and notify the server to be get killed. Insert row 5
and wait for notification from dump thread. Kill the server. This ensures
that both 4 and 5 have reached the semi-sync slave.
Added a new test case:
Insert two rows on master such that first is present in master's binlog and
reached semi sync slave. Second insert should be flushed to binlog but not
sent to slave. Now crash and fail over to slave. The promoted master will send
the extra transaction to slave.
213 lines
7.2 KiB
Plaintext
213 lines
7.2 KiB
Plaintext
# ==== References ====
|
|
#
|
|
# MDEV-21117: recovery for --rpl-semi-sync-slave-enabled server
|
|
#
|
|
|
|
--source include/have_innodb.inc
|
|
--source include/have_debug.inc
|
|
--source include/have_debug_sync.inc
|
|
--source include/have_binlog_format_row.inc
|
|
--source include/master-slave.inc
|
|
|
|
# Initial slave
|
|
--connection server_2
|
|
--source include/stop_slave.inc
|
|
|
|
# Initial master
|
|
--connection server_1
|
|
RESET MASTER;
|
|
SET @@global.max_binlog_size= 4096;
|
|
|
|
--connection server_2
|
|
RESET MASTER;
|
|
SET @@global.max_binlog_size= 4096;
|
|
set @@global.rpl_semi_sync_slave_enabled = 1;
|
|
set @@global.gtid_slave_pos = "";
|
|
CHANGE MASTER TO master_use_gtid= slave_pos;
|
|
--source include/start_slave.inc
|
|
|
|
--connection server_1
|
|
ALTER TABLE mysql.gtid_slave_pos ENGINE=InnoDB;
|
|
set @@global.rpl_semi_sync_master_enabled = 1;
|
|
set @@global.rpl_semi_sync_master_wait_point=AFTER_SYNC;
|
|
|
|
CREATE TABLE t1 (a INT PRIMARY KEY, b MEDIUMTEXT) ENGINE=Innodb;
|
|
INSERT INTO t1 VALUES (1, 'dummy1');
|
|
--save_master_pos
|
|
|
|
--connection server_2
|
|
--sync_with_master
|
|
|
|
--connection server_1
|
|
--let $case = 1
|
|
--echo #
|
|
--echo # Case:$case
|
|
--echo #
|
|
--echo # CRASH the original master, and FAILOVER to the new
|
|
# value 1 for server id 1 -> 2 failover
|
|
--let $failover_to_slave=1
|
|
--let $query_to_crash= INSERT INTO t1 VALUES (2, REPEAT("x", 4100))
|
|
--echo # $query_to_crash
|
|
--echo # Row - 2 will be in master's binlog but not committed, gets replicated
|
|
--echo # to slave and applied. On crash master should have 1 row and slave
|
|
--echo # should have 2 rows.
|
|
--echo #
|
|
--echo # Expected State post crash:
|
|
--echo #=================================================================
|
|
--echo # Master | Slave |
|
|
--echo # 0-1-4 (Not committed) | 0-1-4 (Received through semi-sync |
|
|
--echo # | replication and applied) |
|
|
--echo #=================================================================
|
|
--let $log_search_pattern=truncated binlog file:.*master.*000001
|
|
--let $expected_rows_on_master= 1
|
|
--let $expected_rows_on_slave= 2
|
|
--source rpl_semi_sync_crash.inc
|
|
|
|
--echo #
|
|
--echo # Server_2 promoted as master will send 0-1-4 to new slave Server_1
|
|
--echo #
|
|
--connection server_2
|
|
--let $rows_so_far=3
|
|
--eval INSERT INTO t1 VALUES ($rows_so_far, 'dummy3')
|
|
--save_master_pos
|
|
--echo # The gtid state on current master must be equal to ...
|
|
SHOW VARIABLES LIKE 'gtid_binlog_pos';
|
|
SHOW VARIABLES LIKE 'gtid_slave_pos';
|
|
|
|
--connection server_1
|
|
--sync_with_master
|
|
--eval SELECT COUNT(*) = $rows_so_far as 'true' FROM t1
|
|
--echo # ... the gtid states on the slave:
|
|
SHOW VARIABLES LIKE 'gtid_slave_pos';
|
|
SHOW VARIABLES LIKE 'gtid_binlog_pos';
|
|
|
|
--connection server_2
|
|
--let $case = 2
|
|
--echo #
|
|
--echo # Case:$case
|
|
--echo #
|
|
--echo # CRASH the new master, and FAILOVER back to the original
|
|
# value 0 for the reverse server id 2 -> 1 failover
|
|
--let $failover_to_slave=0
|
|
--let $query_to_crash = INSERT INTO t1 VALUES (4, REPEAT("x", 4100))
|
|
--let $query2_to_crash= INSERT INTO t1 VALUES (5, REPEAT("x", 4100))
|
|
--echo # $query_to_crash
|
|
--echo # $query2_to_crash
|
|
--echo # Rows 4 and 5 will be in master's binlog but not committed, they get
|
|
--echo # replicated to slave and applied. On crash master should have 3 rows
|
|
--echo # and slave should have 5 rows.
|
|
--echo #
|
|
--echo # Expected State post crash:
|
|
--echo #=================================================================
|
|
--echo # Master | Slave |
|
|
--echo # 0-2-6 (Not commited) | 0-2-6 (Received through semi-sync |
|
|
--echo # | replication and applied) |
|
|
--echo # 0-2-7 (Not commited) | 0-2-7 (Received through semi-sync |
|
|
--echo # | replication and applied) |
|
|
--echo #=================================================================
|
|
--let $log_search_pattern=truncated binlog file:.*slave.*000002
|
|
--let $expected_rows_on_master= 3
|
|
--let $expected_rows_on_slave= 5
|
|
--source rpl_semi_sync_crash.inc
|
|
|
|
--echo #
|
|
--echo # Server_1 promoted as master will send 0-2-6 and 0-2-7 to slave Server_2
|
|
--echo #
|
|
--connection server_1
|
|
--let $rows_so_far=6
|
|
--eval INSERT INTO t1 VALUES ($rows_so_far, 'dummy6')
|
|
--save_master_pos
|
|
--echo # The gtid state on current master must be equal to ...
|
|
SHOW VARIABLES LIKE 'gtid_binlog_pos';
|
|
SHOW VARIABLES LIKE 'gtid_slave_pos';
|
|
|
|
--connection server_2
|
|
--sync_with_master
|
|
--eval SELECT COUNT(*) = $rows_so_far as 'true' FROM t1
|
|
--echo # ... the gtid states on the slave:
|
|
SHOW VARIABLES LIKE 'gtid_slave_pos';
|
|
SHOW VARIABLES LIKE 'gtid_binlog_pos';
|
|
|
|
--let $diff_tables=server_1:t1, server_2:t1
|
|
--source include/diff_tables.inc
|
|
|
|
--connection server_1
|
|
--let $case = 3
|
|
--echo #
|
|
--echo # Case:$case
|
|
--echo #
|
|
--echo # CRASH the master and FAILOVER to slave
|
|
--let $failover_to_slave=1
|
|
--let $query_to_crash = INSERT INTO t1 VALUES (7, REPEAT("x", 4100))
|
|
--let $query2_to_crash= INSERT INTO t1 VALUES (8, REPEAT("x", 4100))
|
|
--echo # $query_to_crash
|
|
--echo # $query2_to_crash
|
|
--echo # Rows 7 and 8 will be in master's binlog but not committed, only 7
|
|
--echo # gets replicated to slave and applied. On crash master should have 6
|
|
--echo # rows and slave should have 7 rows.
|
|
--echo #
|
|
--echo # Expected State post crash:
|
|
--echo #=================================================================
|
|
--echo # Master | Slave |
|
|
--echo # 0-1-9 (Not commited) | 0-1-9 (Received through semi-sync |
|
|
--echo # | replication and applied) |
|
|
--echo # 0-1-10 (Not commited - | |
|
|
--echo # never sent to slave) | |
|
|
--echo #=================================================================
|
|
--let $log_search_pattern=truncated binlog file:.*master.*000003
|
|
--let $expected_rows_on_master= 6
|
|
--let $expected_rows_on_slave= 7
|
|
--source rpl_semi_sync_crash.inc
|
|
|
|
--echo #
|
|
--echo # Server_2 promoted as master will send 0-1-9 to slave Server_1
|
|
--echo #
|
|
--connection server_2
|
|
--let $rows_so_far=8
|
|
--eval INSERT INTO t1 VALUES ($rows_so_far, 'Done')
|
|
--source include/save_master_gtid.inc
|
|
--echo # The gtid state on current master must be equal to ...
|
|
SHOW VARIABLES LIKE 'gtid_binlog_pos';
|
|
SHOW VARIABLES LIKE 'gtid_slave_pos';
|
|
|
|
--connection server_1
|
|
--source include/sync_with_master_gtid.inc
|
|
--eval SELECT COUNT(*) = $rows_so_far as 'true' FROM t1
|
|
--echo # ... the gtid states on the slave:
|
|
SHOW VARIABLES LIKE 'gtid_slave_pos';
|
|
SHOW VARIABLES LIKE 'gtid_binlog_pos';
|
|
|
|
--echo #
|
|
--echo # Cleanup
|
|
--echo #
|
|
|
|
--source include/stop_slave.inc
|
|
set global rpl_semi_sync_slave_enabled = 0;
|
|
set global rpl_semi_sync_master_enabled = 0;
|
|
set global rpl_semi_sync_master_wait_point=default;
|
|
RESET MASTER;
|
|
RESET SLAVE;
|
|
|
|
--connection server_2
|
|
RESET MASTER;
|
|
RESET SLAVE;
|
|
set @@global.rpl_semi_sync_master_enabled = 0;
|
|
set @@global.rpl_semi_sync_slave_enabled = 0;
|
|
set @@global.rpl_semi_sync_master_wait_point=default;
|
|
evalp CHANGE MASTER TO master_host='127.0.0.1', master_port=$SERVER_MYPORT_1, master_user='root', master_use_gtid=SLAVE_POS;
|
|
set @@global.gtid_slave_pos=@@global.gtid_binlog_pos;
|
|
--source include/start_slave.inc
|
|
|
|
--connection server_1
|
|
DROP TABLE t1;
|
|
--save_master_pos
|
|
|
|
--connection server_2
|
|
--sync_with_master
|
|
|
|
connection default;
|
|
--enable_reconnect
|
|
--source include/wait_until_connected_again.inc
|
|
|
|
--source include/rpl_end.inc
|