database/mariadb
1
0
Fork 0
mirror of https://github.com/MariaDB/server.git synced 2025-08-07 00:04:31 +03:00
Code Activity

MDEV-21117: refine the server binlog-based recovery for semisync

Browse Source 
Problem:
=======
When the semisync master is crashed and restarted as slave it could
recover transactions that former slaves may never have seen.
A known method existed to clear out all prepared transactions
with --tc-heuristic-recover=rollback does not care to adjust
binlog accordingly.

Fix:
===
The binlog-based recovery is made to concern of the slave semisync role of
post-crash restarted server.
No changes in behavior is done to the "normal" binloggging server
and the semisync master.

When the restarted server is configured with
  --rpl-semi-sync-slave-enabled=1
the refined recovery attempts to roll back prepared transactions
and truncate binlog accordingly.
In case of a partially committed (that is committed at least
in one of the engine participants) such transaction gets committed.
It's guaranteed no (partially as well) committed transactions
exist beyond the truncate position.
In case there exists a non-transactional replication event
(being in a way a committed transaction) past the
computed truncate position the recovery ends with an error.

As after master crash and failover to slave, the demoted-to-slave
ex-master must be ready to face and accept its own (generated by)
events, without generally necessary --replicate-same-server-id.
So the acceptance conditions are relaxed for the semisync slave
to accept own events without that option.
While gtid_strict_mode ON ensures no duplicate transaction can be
(re-)executed the master_use_gtid=none slave has to be
configured with --replicate-same-server-id.

*NOTE* for reviewers.

This patch does not handle the user XA which is done
in next git commit.
This commit is contained in:
Sujatha
2020-04-09 20:45:45 +05:30
committed by Andrei Elkin
parent 82c07b178a
commit 6c39eaeb12
25 changed files with 2619 additions and 127 deletions
Download Patch File Download Diff File Expand all files Collapse all files

189
mysql-test/suite/binlog/r/binlog_truncate_multi_engine.result Normal file
View File

@@ -0,0 +1,189 @@
call mtr.add_suppression("Can.t init tc log");
call mtr.add_suppression("Aborting");
CREATE TABLE t1 (a INT PRIMARY KEY, b MEDIUMTEXT) ENGINE=Innodb;
CREATE TABLE t2 (a INT PRIMARY KEY, b MEDIUMTEXT) ENGINE=rocksdb;
#
#
# Case "A" : "neither engine committed => rollback & binlog truncate"
#
RESET MASTER;
FLUSH LOGS;
SET GLOBAL max_binlog_size= 4096;
connect con1,localhost,root,,;
List of binary logs before rotation
show binary logs;
Log_name File_size
master-bin.000001 #
master-bin.000002 #
INSERT INTO t1 VALUES (1, REPEAT("x", 1));
INSERT INTO t2 VALUES (1, REPEAT("x", 1));
BEGIN;
INSERT INTO t2 VALUES (2, REPEAT("x", 4100));
INSERT INTO t1 VALUES (2, REPEAT("x", 4100));
SET DEBUG_SYNC= "commit_after_release_LOCK_log SIGNAL con1_ready WAIT_FOR signal_no_signal";
COMMIT;
connection default;
SET DEBUG_SYNC= "now WAIT_FOR con1_ready";
List of binary logs after rotation
show binary logs;
Log_name File_size
master-bin.000001 #
master-bin.000002 #
master-bin.000003 #
# restart the server with --rpl-semi-sync-slave-enabled=1
# the server is restarted
# restart: --rpl-semi-sync-slave-enabled=1
connection default;
#
# *** Summary: 1 row should be present in both tables; binlog is truncated; number of binlogs at reconnect - 3:
#
SELECT COUNT(*) FROM t1;
COUNT(*)
1
SELECT COUNT(*) FROM t2;
COUNT(*)
1
SELECT @@GLOBAL.gtid_binlog_state;
@@GLOBAL.gtid_binlog_state
0-1-2
SELECT @@GLOBAL.gtid_binlog_pos;
@@GLOBAL.gtid_binlog_pos
0-1-2
List of binary logs at the end of the tests
show binary logs;
Log_name File_size
master-bin.000001 #
master-bin.000002 #
master-bin.000003 #
# ***
DELETE FROM t1;
DELETE FROM t2;
disconnect con1;
#
Proof of the truncated binlog file is readable (two transactions must be seen):
/*!50530 SET @@SESSION.PSEUDO_SLAVE_MODE=1*/;
/*!40019 SET @@session.max_insert_delayed_threads=0*/;
/*!50003 SET @OLD_COMPLETION_TYPE=@@COMPLETION_TYPE,COMPLETION_TYPE=0*/;
DELIMITER /*!*/;
START TRANSACTION
/*!*/;
COMMIT/*!*/;
START TRANSACTION
/*!*/;
COMMIT/*!*/;
DELIMITER ;
# End of log file
ROLLBACK /* added by mysqlbinlog */;
/*!50003 SET COMPLETION_TYPE=@OLD_COMPLETION_TYPE*/;
/*!50530 SET @@SESSION.PSEUDO_SLAVE_MODE=0*/;
#
#
# Case "B" : "one engine has committed its transaction branch"
#
RESET MASTER;
FLUSH LOGS;
SET GLOBAL max_binlog_size= 4096;
connect con1,localhost,root,,;
List of binary logs before rotation
show binary logs;
Log_name File_size
master-bin.000001 #
master-bin.000002 #
INSERT INTO t1 VALUES (1, REPEAT("x", 1));
INSERT INTO t2 VALUES (1, REPEAT("x", 1));
SET GLOBAL debug_dbug="d,enable_log_write_upto_crash";
BEGIN;
INSERT INTO t2 VALUES (2, REPEAT("x", 4100));
INSERT INTO t1 VALUES (2, REPEAT("x", 4100));
COMMIT;
connection default;
# restart: --rpl-semi-sync-slave-enabled=1
connection default;
#
# *** Summary: 2 rows should be present in both tables; no binlog truncation; one extra binlog file compare with A; number of binlogs at reconnect - 4:
#
SELECT COUNT(*) FROM t1;
COUNT(*)
2
SELECT COUNT(*) FROM t2;
COUNT(*)
2
SELECT @@GLOBAL.gtid_binlog_state;
@@GLOBAL.gtid_binlog_state
0-1-3
SELECT @@GLOBAL.gtid_binlog_pos;
@@GLOBAL.gtid_binlog_pos
0-1-3
List of binary logs at the end of the tests
show binary logs;
Log_name File_size
master-bin.000001 #
master-bin.000002 #
master-bin.000003 #
master-bin.000004 #
# ***
DELETE FROM t1;
DELETE FROM t2;
disconnect con1;
#
#
#
# Case "C" : "both engines have committed its transaction branch"
#
RESET MASTER;
FLUSH LOGS;
SET GLOBAL max_binlog_size= 4096;
connect con1,localhost,root,,;
List of binary logs before rotation
show binary logs;
Log_name File_size
master-bin.000001 #
master-bin.000002 #
INSERT INTO t1 VALUES (1, REPEAT("x", 1));
INSERT INTO t2 VALUES (1, REPEAT("x", 1));
BEGIN;
INSERT INTO t2 VALUES (2, REPEAT("x", 4100));
INSERT INTO t1 VALUES (2, REPEAT("x", 4100));
SET DEBUG_SYNC= "commit_after_run_commit_ordered SIGNAL con1_ready WAIT_FOR signal_no_signal";
COMMIT;
connection default;
SET DEBUG_SYNC= "now WAIT_FOR con1_ready";
List of binary logs after rotation
show binary logs;
Log_name File_size
master-bin.000001 #
master-bin.000002 #
master-bin.000003 #
# restart the server with --rpl-semi-sync-slave-enabled=1
# the server is restarted
# restart: --rpl-semi-sync-slave-enabled=1
connection default;
#
# *** Summary: 2 rows should be present in both tables; no binlog truncation; the same # of binlog files as in B; number of binlogs at reconnect - 4:
#
SELECT COUNT(*) FROM t1;
COUNT(*)
2
SELECT COUNT(*) FROM t2;
COUNT(*)
2
SELECT @@GLOBAL.gtid_binlog_state;
@@GLOBAL.gtid_binlog_state
0-1-3
SELECT @@GLOBAL.gtid_binlog_pos;
@@GLOBAL.gtid_binlog_pos
0-1-3
List of binary logs at the end of the tests
show binary logs;
Log_name File_size
master-bin.000001 #
master-bin.000002 #
master-bin.000003 #
master-bin.000004 #
# ***
DELETE FROM t1;
DELETE FROM t2;
disconnect con1;
#
DROP TABLE t1, t2;
# End of the tests