mariadb/mysql-test/suite/rpl/r/rpl_parallel_sbm.result
Brandon Nesterenko 68938d2b42 MDEV-33500 (part 2): rpl.rpl_parallel_sbm can still fail
The failing test case validates Seconds_Behind_Master for a delayed
slave, while STOP SLAVE is executed during a delay. The test fixes
initially added to the test (commit b04c857596) added a table lock
to ensure a transaction could not finish before validating the
Seconds_Behind_Master field after SLAVE START, but did not address a
possibility that the transaction could finish before running the
STOP SLAVE command, which invalidates the validations for the rest
of the test case. Specifically, this would result in 1) a timeout in
“Waiting for table metadata lock” on the replica, which expects the
transaction to retry after slave restart and hit a lock conflict on
the locked tables (added in b04c857596), and 2) that
Seconds_Behind_Master should have increased, but did not.

The failure can be reproduced by synchronizing the slave to the master
before the MDEV-32265 echo statement (i.e. before the SLAVE STOP).

This patch fixes the test by adding a mechanism to use DEBUG_SYNC to
synchronize a MASTER_DELAY, rather than continually increase the
duration of the delay each time the test fails on buildbot. This is
to ensure that on slow machines, a delay does not pass before the
test gets a chance to validate results. Additionally, it decreases
overall test time because the test can continue immediately after
validation, thereby bypassing the remainder of a full delay for each
transaction.
2024-09-17 06:29:20 -06:00

131 lines
4.6 KiB
Text

include/master-slave.inc
[connection master]
#
# MDEV-29639: Seconds_Behind_Master is incorrect for Delayed, Parallel Replicas
#
connection slave;
include/stop_slave.inc
set @@GLOBAL.slave_parallel_mode= CONSERVATIVE;
set @@GLOBAL.debug_dbug= "d,negate_clock_diff_with_master,sql_delay_by_debug_sync";
change master to master_delay=1, master_use_gtid=Slave_Pos;
include/start_slave.inc
connection master;
create table t1 (a int);
create table t2 (a int);
include/save_master_gtid.inc
connection slave;
include/sync_with_master_sql_delay_debug_sync.inc
set debug_sync= "now WAIT_FOR at_sql_delay";
set debug_sync= "now SIGNAL continue_sql_thread";
set debug_sync= "now WAIT_FOR at_sql_delay";
set debug_sync= "now SIGNAL continue_sql_thread";
#
# Pt 1) Ensure SBM is updated immediately upon arrival of the next event
connection master;
# Sleep 2 to allow a buffer between events for SBM check
insert into t1 values (0);
include/save_master_gtid.inc
connection slave;
# Waiting for transaction to arrive on slave and begin SQL Delay..
set debug_sync= "now WAIT_FOR at_sql_delay";
# Validating SBM is updated on event arrival..
# ..done
# MDEV-32265. At time of STOP SLAVE, if the SQL Thread is currently
# delaying a transaction; then when the reciprocal START SLAVE occurs,
# if the event is still to be delayed, SBM should resume accordingly
connection server_2;
# Ensure the kill from STOP SLAVE will be received before continuing the
# SQL thread
set debug_sync="after_thd_awake_kill SIGNAL slave_notified_of_kill";
STOP SLAVE;
connection slave;
set debug_sync= "now WAIT_FOR slave_notified_of_kill";
set debug_sync= "now SIGNAL continue_sql_thread";
connection server_2;
include/wait_for_slave_to_stop.inc
set debug_sync="RESET";
# Lock t1 on slave to ensure the event can't finish (and thereby update
# Seconds_Behind_Master) so slow running servers don't accidentally
# catch up to the master before checking SBM.
connection server_2;
LOCK TABLES t1 WRITE;
include/start_slave.inc
connection slave;
# SQL delay has no impact for the rest of the test case, so ignore it
include/sync_with_master_sql_delay_debug_sync.inc
set debug_sync= "now WAIT_FOR at_sql_delay";
set debug_sync= "now SIGNAL continue_sql_thread";
# Waiting for replica to get blocked by the table lock
# Sleeping 1s to increment SBM
# Ensuring Seconds_Behind_Master increases after sleeping..
# ..done
connection server_2;
UNLOCK TABLES;
include/sync_with_master_gtid.inc
#
# Pt 2) If the worker threads have not entered an idle state, ensure
# following events do not update SBM
connection slave;
LOCK TABLES t1 WRITE;
connection master;
# Sleep 2 to allow a buffer between events for SBM check
insert into t1 values (1);
# Sleep 3 to create gap between events
insert into t1 values (2);
include/save_master_pos.inc
connection slave;
# Wait for first transaction to complete SQL delay and begin execution..
include/sync_with_master_sql_delay_debug_sync.inc
set debug_sync= "now WAIT_FOR at_sql_delay";
set debug_sync= "now SIGNAL continue_sql_thread";
# Wait for second transaction to complete SQL delay..
include/sync_with_master_sql_delay_debug_sync.inc
set debug_sync= "now WAIT_FOR at_sql_delay";
set debug_sync= "now SIGNAL continue_sql_thread";
# Validate SBM calculation doesn't use the second transaction because worker threads shouldn't have gone idle..
# ..and that SBM wasn't calculated using prior committed transactions
# ..done
connection slave;
UNLOCK TABLES;
include/wait_for_slave_param.inc [Relay_Master_Log_File]
include/wait_for_slave_param.inc [Exec_Master_Log_Pos]
# Cleanup
include/stop_slave.inc
set debug_sync= "RESET";
set @@GLOBAL.debug_dbug= "-d,sql_delay_by_debug_sync";
CHANGE MASTER TO master_delay=0;
include/start_slave.inc
#
# MDEV-30619: Parallel Slave SQL Thread Can Update Seconds_Behind_Master with Active Workers
#
connection slave;
# Ensure the replica is fully idle before starting transactions
# Lock t1 on slave so the first received transaction does not complete/commit
LOCK TABLES t1 WRITE;
connection master;
insert into t1 values (3);
include/save_master_gtid.inc
connection slave;
# Waiting for first transaction to begin..
connection master;
# Sleep 2 sec to create a gap between events
INSERT INTO t2 VALUES (1);
include/save_master_gtid.inc
connection slave;
# Waiting for second transaction to begin..
connection slave;
UNLOCK TABLES;
include/sync_with_master_gtid.inc
#
# Cleanup
connection master;
DROP TABLE t1, t2;
include/save_master_gtid.inc
connection slave;
include/sync_with_master_gtid.inc
include/stop_slave.inc
set @@GLOBAL.debug_dbug= "";
set @@GLOBAL.slave_parallel_mode= "$save_parallel_mode";
include/start_slave.inc
include/rpl_end.inc
# End of rpl_parallel_sbm.test