database/mariadb
1
0
Fork 0
mirror of https://github.com/MariaDB/server.git synced 2025-11-30 05:23:50 +03:00
Code Activity
Files
ded27d28968e6e47e901b9472001550ec3dbfdd1
mariadb/mysql-test/suite/rpl/include/rpl_parallel_slave_bgc_kill.inc
Marko Mäkelä 3cef4f8f0f MDEV-515 Reduce InnoDB undo logging for insert into empty table 
We implement an idea that was suggested by Michael 'Monty' Widenius
in October 2017: When InnoDB is inserting into an empty table or partition,
we can write a single undo log record TRX_UNDO_EMPTY, which will cause
ROLLBACK to clear the table.

For this to work, the insert into an empty table or partition must be
covered by an exclusive table lock that will be held until the transaction
has been committed or rolled back, or the INSERT operation has been
rolled back (and the table is empty again), in lock_table_x_unlock().

Clustered index records that are covered by the TRX_UNDO_EMPTY record
will carry DB_TRX_ID=0 and DB_ROLL_PTR=1<<55, and thus they cannot
be distinguished from what MDEV-12288 leaves behind after purging the
history of row-logged operations.

Concurrent non-locking reads must be adjusted: If the read view was
created before the INSERT into an empty table, then we must continue
to imagine that the table is empty, and not try to read any records.
If the read view was created after the INSERT was committed, then
all records must be visible normally. To implement this, we introduce
the field dict_table_t::bulk_trx_id.

This special handling only applies to the very first INSERT statement
of a transaction for the empty table or partition. If a subsequent
statement in the transaction is modifying the initially empty table again,
we must enable row-level undo logging, so that we will be able to
roll back to the start of the statement in case of an error (such as
duplicate key).

INSERT IGNORE will continue to use row-level logging and locking, because
implementing it would require the ability to roll back the latest row.
Since the undo log that we write only allows us to roll back the entire
statement, we cannot support INSERT IGNORE. We will introduce a
handler::extra() parameter HA_EXTRA_IGNORE_INSERT to indicate to storage
engines that INSERT IGNORE is being executed.

In many test cases, we add an extra record to the table, so that during
the 'interesting' part of the test, row-level locking and logging will
be used.

Replicas will continue to use row-level logging and locking until
MDEV-24622 has been addressed. Likewise, this optimization will be
disabled in Galera cluster until MDEV-24623 enables it.

dict_table_t::bulk_trx_id: The latest active or committed transaction
that initiated an insert into an empty table or partition.
Protected by exclusive table lock and a clustered index leaf page latch.

ins_node_t::bulk_insert: Whether bulk insert was initiated.

trx_t::mod_tables: Use C++11 style accessors (emplace instead of insert).
Unlike earlier, this collection will cover also temporary tables.

trx_mod_table_time_t: Add start_bulk_insert(), end_bulk_insert(),
is_bulk_insert(), was_bulk_insert().

trx_undo_report_row_operation(): Before accessing any undo log pages,
invoke trx->mod_tables.emplace() in order to determine whether undo
logging was disabled, or whether this is the first INSERT and we are
supposed to write a TRX_UNDO_EMPTY record.

row_ins_clust_index_entry_low(): If we are inserting into an empty
clustered index leaf page, set the ins_node_t::bulk_insert flag for
the subsequent trx_undo_report_row_operation() call.

lock_rec_insert_check_and_lock(), lock_prdt_insert_check_and_lock():
Remove the redundant parameter 'flags' that can be checked in the caller.

btr_cur_ins_lock_and_undo(): Simplify the logic. Correctly write
DB_TRX_ID,DB_ROLL_PTR after invoking trx_undo_report_row_operation().

trx_mark_sql_stat_end(), ha_innobase::extra(HA_EXTRA_IGNORE_INSERT),
ha_innobase::external_lock(): Invoke trx_t::end_bulk_insert() so that
the next statement will not be covered by table-level undo logging.

ReadView::changes_visible(trx_id_t) const: New accessor for the case
where the trx_id_t is not read from a potentially corrupted index page
but directly from the memory. In this case, we can skip a sanity check.

row_sel(), row_sel_try_search_shortcut(), row_search_mvcc():
row_sel_try_search_shortcut_for_mysql(),
row_merge_read_clustered_index(): Check dict_table_t::bulk_trx_id.

row_sel_clust_sees(): Replaces lock_clust_rec_cons_read_sees().

lock_sec_rec_cons_read_sees(): Replaced with lower-level code.

btr_root_page_init(): Refactored from btr_create().

dict_index_t::clear(), dict_table_t::clear(): Empty an index or table,
for the ROLLBACK of an INSERT operation.

ROW_T_EMPTY, ROW_OP_EMPTY: Note a concurrent ROLLBACK of an INSERT
into an empty table.

This is joint work with Thirunarayanan Balathandayuthapani,
who created a working prototype.
Thanks to Matthias Leich for extensive testing.
2021-01-25 18:41:27 +02:00

459 lines
13 KiB
SQL
Raw Blame History

--echo *** Test killing slave threads at various wait points ***
--source include/have_innodb.inc
--source include/have_debug.inc
--source include/have_debug_sync.inc
--source include/have_binlog_format_statement.inc
--source include/master-slave.inc
# Test various aspects of parallel replication.
--connection server_2
SET @old_parallel_threads=@@GLOBAL.slave_parallel_threads;
SET @old_parallel_mode=@@GLOBAL.slave_parallel_mode;
--source include/stop_slave.inc
SET GLOBAL slave_parallel_threads=10;
SET GLOBAL slave_parallel_mode='conservative';
CHANGE MASTER TO master_use_gtid=slave_pos;
--source include/start_slave.inc
--connection server_1
--connect (con_temp3,127.0.0.1,root,,test,$SERVER_MYPORT_1,)
--connect (con_temp4,127.0.0.1,root,,test,$SERVER_MYPORT_1,)
--connect (con_temp5,127.0.0.1,root,,test,$SERVER_MYPORT_1,)
ALTER TABLE mysql.gtid_slave_pos ENGINE=InnoDB;
CREATE TABLE t1 (a int PRIMARY KEY) ENGINE=MyISAM;
CREATE TABLE t2 (a int PRIMARY KEY) ENGINE=InnoDB;
CREATE TABLE t3 (a INT PRIMARY KEY, b INT) ENGINE=InnoDB;
# MDEV-515 takes X-lock on the table for the first insert.
# So concurrent insert won't happen on the table
INSERT INTO t2 VALUES(100);
INSERT INTO t3 VALUES(100, 100);
--save_master_pos
--connection server_2
--sync_with_master
--connection server_1
# Use a stored function to inject a debug_sync into the appropriate THD.
# The function does nothing on the master, and on the slave it injects the
# desired debug_sync action(s).
SET sql_log_bin=0;
--delimiter ||
CREATE FUNCTION foo(x INT, d1 VARCHAR(500), d2 VARCHAR(500))
RETURNS INT DETERMINISTIC
BEGIN
RETURN x;
END
||
--delimiter ;
SET sql_log_bin=1;
--connection server_2
SET sql_log_bin=0;
--delimiter ||
CREATE FUNCTION foo(x INT, d1 VARCHAR(500), d2 VARCHAR(500))
RETURNS INT DETERMINISTIC
BEGIN
IF d1 != '' THEN
SET debug_sync = d1;
END IF;
IF d2 != '' THEN
SET debug_sync = d2;
END IF;
RETURN x;
END
||
--delimiter ;
SET sql_log_bin=1;
--echo *** 1. Test killing transaction waiting in commit for previous transaction to commit ***
# Set up three transactions on the master that will be group-committed
# together so they can be replicated in parallel on the slave.
--connection con_temp3
SET debug_sync='commit_after_release_LOCK_prepare_ordered SIGNAL master_queued1 WAIT_FOR master_cont1';
send INSERT INTO t3 VALUES (31, foo(31,
'commit_before_prepare_ordered WAIT_FOR t2_waiting',
'commit_after_prepare_ordered SIGNAL t1_ready WAIT_FOR t1_cont'));
--connection server_1
SET debug_sync='now WAIT_FOR master_queued1';
--connection con_temp4
SET debug_sync='commit_after_release_LOCK_prepare_ordered SIGNAL master_queued2';
BEGIN;
# This insert is just so we can get T2 to wait while a query is running that we
# can see in SHOW PROCESSLIST so we can get its thread_id to kill later.
INSERT INTO t3 VALUES (32, foo(32,
'ha_write_row_end SIGNAL t2_query WAIT_FOR t2_cont',
''));
# This insert sets up debug_sync points so that T2 will tell when it is at its
# wait point where we want to kill it - and when it has been killed.
INSERT INTO t3 VALUES (33, foo(33,
'group_commit_waiting_for_prior SIGNAL t2_waiting',
'group_commit_waiting_for_prior_killed SIGNAL t2_killed'));
send COMMIT;
--connection server_1
SET debug_sync='now WAIT_FOR master_queued2';
--connection con_temp5
SET debug_sync='commit_after_release_LOCK_prepare_ordered SIGNAL master_queued3';
send INSERT INTO t3 VALUES (34, foo(34,
'',
''));
--connection server_1
SET debug_sync='now WAIT_FOR master_queued3';
SET debug_sync='now SIGNAL master_cont1';
--connection con_temp3
REAP;
--connection con_temp4
REAP;
--connection con_temp5
REAP;
--connection server_1
SELECT * FROM t3 WHERE a >= 30 ORDER BY a;
SET debug_sync='RESET';
--connection server_2
SET sql_log_bin=0;
CALL mtr.add_suppression("Query execution was interrupted");
CALL mtr.add_suppression("Commit failed due to failure of an earlier commit on which this one depends");
CALL mtr.add_suppression("Slave: Connection was killed");
SET sql_log_bin=1;
# Wait until T2 is inside executing its insert of 32, then find it in SHOW
# PROCESSLIST to know its thread id for KILL later.
SET debug_sync='now WAIT_FOR t2_query';
--let $thd_id= `SELECT ID FROM INFORMATION_SCHEMA.PROCESSLIST WHERE INFO LIKE '%foo(32%' AND INFO NOT LIKE '%LIKE%'`
SET debug_sync='now SIGNAL t2_cont';
# Wait until T2 has entered its wait for T1 to commit, and T1 has
# progressed into its commit phase.
SET debug_sync='now WAIT_FOR t1_ready';
# Now kill the transaction T2.
--replace_result $thd_id THD_ID
eval KILL $thd_id;
# Wait until T2 has reacted on the kill.
SET debug_sync='now WAIT_FOR t2_killed';
# Now we can allow T1 to proceed.
SET debug_sync='now SIGNAL t1_cont';
--let $slave_sql_errno= 1317,1927,1964
--source include/wait_for_slave_sql_error.inc
STOP SLAVE IO_THREAD;
SELECT * FROM t3 WHERE a >= 30 ORDER BY a;
# Now we have to disable the debug_sync statements, so they do not trigger
# when the events are retried.
SET debug_sync='RESET';
SET GLOBAL slave_parallel_threads=0;
SET GLOBAL slave_parallel_threads=10;
SET sql_log_bin=0;
DROP FUNCTION foo;
--delimiter ||
CREATE FUNCTION foo(x INT, d1 VARCHAR(500), d2 VARCHAR(500))
RETURNS INT DETERMINISTIC
BEGIN
RETURN x;
END
||
--delimiter ;
SET sql_log_bin=1;
--connection server_1
INSERT INTO t3 VALUES (39,0);
--save_master_pos
--connection server_2
--source include/start_slave.inc
--sync_with_master
SELECT * FROM t3 WHERE a >= 30 ORDER BY a;
# Restore the foo() function.
SET sql_log_bin=0;
DROP FUNCTION foo;
--delimiter ||
CREATE FUNCTION foo(x INT, d1 VARCHAR(500), d2 VARCHAR(500))
RETURNS INT DETERMINISTIC
BEGIN
IF d1 != '' THEN
SET debug_sync = d1;
END IF;
IF d2 != '' THEN
SET debug_sync = d2;
END IF;
RETURN x;
END
||
--delimiter ;
SET sql_log_bin=1;
--connection server_2
# Respawn all worker threads to clear any left-over debug_sync or other stuff.
--source include/stop_slave.inc
SET GLOBAL slave_parallel_threads=0;
SET GLOBAL slave_parallel_threads=10;
--source include/start_slave.inc
--echo *** 2. Same as (1), but without restarting IO thread after kill of SQL threads ***
# Set up three transactions on the master that will be group-committed
# together so they can be replicated in parallel on the slave.
--connection con_temp3
SET debug_sync='commit_after_release_LOCK_prepare_ordered SIGNAL master_queued1 WAIT_FOR master_cont1';
send INSERT INTO t3 VALUES (41, foo(41,
'commit_before_prepare_ordered WAIT_FOR t2_waiting',
'commit_after_prepare_ordered SIGNAL t1_ready WAIT_FOR t1_cont'));
--connection server_1
SET debug_sync='now WAIT_FOR master_queued1';
--connection con_temp4
SET debug_sync='commit_after_release_LOCK_prepare_ordered SIGNAL master_queued2';
BEGIN;
# This insert is just so we can get T2 to wait while a query is running that we
# can see in SHOW PROCESSLIST so we can get its thread_id to kill later.
INSERT INTO t3 VALUES (42, foo(42,
'ha_write_row_end SIGNAL t2_query WAIT_FOR t2_cont',
''));
# This insert sets up debug_sync points so that T2 will tell when it is at its
# wait point where we want to kill it - and when it has been killed.
INSERT INTO t3 VALUES (43, foo(43,
'group_commit_waiting_for_prior SIGNAL t2_waiting',
'group_commit_waiting_for_prior_killed SIGNAL t2_killed'));
send COMMIT;
--connection server_1
SET debug_sync='now WAIT_FOR master_queued2';
--connection con_temp5
SET debug_sync='commit_after_release_LOCK_prepare_ordered SIGNAL master_queued3';
send INSERT INTO t3 VALUES (44, foo(44,
'',
''));
--connection server_1
SET debug_sync='now WAIT_FOR master_queued3';
SET debug_sync='now SIGNAL master_cont1';
--connection con_temp3
REAP;
--connection con_temp4
REAP;
--connection con_temp5
REAP;
--connection server_1
SELECT * FROM t3 WHERE a >= 40 ORDER BY a;
SET debug_sync='RESET';
--connection server_2
# Wait until T2 is inside executing its insert of 42, then find it in SHOW
# PROCESSLIST to know its thread id for KILL later.
SET debug_sync='now WAIT_FOR t2_query';
--let $thd_id= `SELECT ID FROM INFORMATION_SCHEMA.PROCESSLIST WHERE INFO LIKE '%foo(42%' AND INFO NOT LIKE '%LIKE%'`
SET debug_sync='now SIGNAL t2_cont';
# Wait until T2 has entered its wait for T1 to commit, and T1 has
# progressed into its commit phase.
SET debug_sync='now WAIT_FOR t1_ready';
# Now kill the transaction T2.
--replace_result $thd_id THD_ID
eval KILL $thd_id;
# Wait until T2 has reacted on the kill.
SET debug_sync='now WAIT_FOR t2_killed';
# Now we can allow T1 to proceed.
SET debug_sync='now SIGNAL t1_cont';
--let $slave_sql_errno= 1317,1927,1964
--source include/wait_for_slave_sql_error.inc
# Now we have to disable the debug_sync statements, so they do not trigger
# when the events are retried.
SET debug_sync='RESET';
SET GLOBAL slave_parallel_threads=0;
SET GLOBAL slave_parallel_threads=10;
SET sql_log_bin=0;
DROP FUNCTION foo;
--delimiter ||
CREATE FUNCTION foo(x INT, d1 VARCHAR(500), d2 VARCHAR(500))
RETURNS INT DETERMINISTIC
BEGIN
RETURN x;
END
||
--delimiter ;
SET sql_log_bin=1;
--connection server_1
INSERT INTO t3 VALUES (49,0);
--save_master_pos
--connection server_2
START SLAVE SQL_THREAD;
--sync_with_master
SELECT * FROM t3 WHERE a >= 40 ORDER BY a;
# Restore the foo() function.
SET sql_log_bin=0;
DROP FUNCTION foo;
--delimiter ||
CREATE FUNCTION foo(x INT, d1 VARCHAR(500), d2 VARCHAR(500))
RETURNS INT DETERMINISTIC
BEGIN
IF d1 != '' THEN
SET debug_sync = d1;
END IF;
IF d2 != '' THEN
SET debug_sync = d2;
END IF;
RETURN x;
END
||
--delimiter ;
SET sql_log_bin=1;
--connection server_2
# Respawn all worker threads to clear any left-over debug_sync or other stuff.
--source include/stop_slave.inc
SET GLOBAL slave_parallel_threads=0;
SET GLOBAL slave_parallel_threads=10;
--source include/start_slave.inc
--echo *** 3. Same as (2), but not using gtid mode ***
--connection server_2
--source include/stop_slave.inc
CHANGE MASTER TO master_use_gtid=no;
--source include/start_slave.inc
--connection server_1
# Set up three transactions on the master that will be group-committed
# together so they can be replicated in parallel on the slave.
--connection con_temp3
SET debug_sync='commit_after_release_LOCK_prepare_ordered SIGNAL master_queued1 WAIT_FOR master_cont1';
send INSERT INTO t3 VALUES (51, foo(51,
'commit_before_prepare_ordered WAIT_FOR t2_waiting',
'commit_after_prepare_ordered SIGNAL t1_ready WAIT_FOR t1_cont'));
--connection server_1
SET debug_sync='now WAIT_FOR master_queued1';
--connection con_temp4
SET debug_sync='commit_after_release_LOCK_prepare_ordered SIGNAL master_queued2';
BEGIN;
# This insert is just so we can get T2 to wait while a query is running that we
# can see in SHOW PROCESSLIST so we can get its thread_id to kill later.
INSERT INTO t3 VALUES (52, foo(52,
'ha_write_row_end SIGNAL t2_query WAIT_FOR t2_cont',
''));
# This insert sets up debug_sync points so that T2 will tell when it is at its
# wait point where we want to kill it - and when it has been killed.
INSERT INTO t3 VALUES (53, foo(53,
'group_commit_waiting_for_prior SIGNAL t2_waiting',
'group_commit_waiting_for_prior_killed SIGNAL t2_killed'));
send COMMIT;
--connection server_1
SET debug_sync='now WAIT_FOR master_queued2';
--connection con_temp5
SET debug_sync='commit_after_release_LOCK_prepare_ordered SIGNAL master_queued3';
send INSERT INTO t3 VALUES (54, foo(54,
'',
''));
--connection server_1
SET debug_sync='now WAIT_FOR master_queued3';
SET debug_sync='now SIGNAL master_cont1';
--connection con_temp3
REAP;
--connection con_temp4
REAP;
--connection con_temp5
REAP;
--connection server_1
SELECT * FROM t3 WHERE a >= 50 ORDER BY a;
SET debug_sync='RESET';
--connection server_2
# Wait until T2 is inside executing its insert of 52, then find it in SHOW
# PROCESSLIST to know its thread id for KILL later.
SET debug_sync='now WAIT_FOR t2_query';
--let $thd_id= `SELECT ID FROM INFORMATION_SCHEMA.PROCESSLIST WHERE INFO LIKE '%foo(52%' AND INFO NOT LIKE '%LIKE%'`
SET debug_sync='now SIGNAL t2_cont';
# Wait until T2 has entered its wait for T1 to commit, and T1 has
# progressed into its commit phase.
SET debug_sync='now WAIT_FOR t1_ready';
# Now kill the transaction T2.
--replace_result $thd_id THD_ID
eval KILL $thd_id;
# Wait until T2 has reacted on the kill.
SET debug_sync='now WAIT_FOR t2_killed';
# Now we can allow T1 to proceed.
SET debug_sync='now SIGNAL t1_cont';
--let $slave_sql_errno= 1317,1927,1964
--source include/wait_for_slave_sql_error.inc
SELECT * FROM t3 WHERE a >= 50 ORDER BY a;
# Now we have to disable the debug_sync statements, so they do not trigger
# when the events are retried.
SET debug_sync='RESET';
SET GLOBAL slave_parallel_threads=0;
SET GLOBAL slave_parallel_threads=10;
SET sql_log_bin=0;
DROP FUNCTION foo;
--delimiter ||
CREATE FUNCTION foo(x INT, d1 VARCHAR(500), d2 VARCHAR(500))
RETURNS INT DETERMINISTIC
BEGIN
RETURN x;
END
||
--delimiter ;
SET sql_log_bin=1;
--connection server_1
INSERT INTO t3 VALUES (59,0);
--save_master_pos
--connection server_2
START SLAVE SQL_THREAD;
--sync_with_master
SELECT * FROM t3 WHERE a >= 50 ORDER BY a;
# Clean up.
--connection server_2
--source include/stop_slave.inc
CHANGE MASTER TO master_use_gtid=slave_pos;
SET GLOBAL slave_parallel_threads=@old_parallel_threads;
SET GLOBAL slave_parallel_mode=@old_parallel_mode;
--source include/start_slave.inc
SET DEBUG_SYNC= 'RESET';
--connection server_1
DROP function foo;
DROP TABLE t1,t2,t3;
SET DEBUG_SYNC= 'RESET';
--source include/rpl_end.inc
View Git Blame Copy Permalink