database/mariadb
1
0
Fork 0
mirror of https://github.com/MariaDB/server.git synced 2025-07-30 16:24:05 +03:00
Code Activity

Bug#11763573 - 56299: MUTEX DEADLOCK WITH COM_BINLOG_DUMP, BINLOG PURGE, AND PROCESSLIST/KILL

Browse Source 
The bug case is similar to one fixed earlier bug_49536.
Deadlock involving LOCK_log appears to be possible because the purge running thread
is holding LOCK_log whereas there is no sense of doing that and which fact was
exploited by the earlier bug fixes.

Fixed with small reengineering of rotate_and_purge(), adding two new methods and
setting up a policy to execute those instead of the former
rotate_and_purge(RP_LOCK_LOG_IS_ALREADY_LOCKED).
The policy for using rotate(), purge() is that if the caller acquires LOCK_log itself,
it should call rotate(), release the mutex and run purge().

Side effect of this patch is refining error message of bug@11747416 to print
the whole path.



mysql-test/suite/rpl/r/rpl_cant_read_event_incident.result:
  the file name printing is changed to a relative path instead of just the file name.
mysql-test/suite/rpl/r/rpl_log_pos.result:
  the file name printing is changed to a relative path instead of just the file name.
mysql-test/suite/rpl/r/rpl_manual_change_index_file.result:
  the file name printing is changed to a relative path instead of just the file name.
mysql-test/suite/rpl/r/rpl_packet.result:
  the file name printing is changed to a relative path instead of just the file name.
mysql-test/suite/rpl/r/rpl_rotate_purge_deadlock.result:
  new result file is added.
mysql-test/suite/rpl/t/rpl_cant_read_event_incident.test:
  The test of that bug can't satisfy windows and unix backslash interpretation so windows
  execution is chosen to bypass.
mysql-test/suite/rpl/t/rpl_rotate_purge_deadlock-master.opt:
  new opt file is added.
mysql-test/suite/rpl/t/rpl_rotate_purge_deadlock.test:
  regression test is added as well as verification of a 
  possible side effect of the fixes is tried.
sql/log.cc:
  LOCK_log is never taken during execution of log purging routine.
  The former MYSQL_BIN_LOG::rotate_and_purge is made to necessarily 
  acquiring and releasing LOCK_log. 
  If caller takes the mutex itself it has to use a new rotate(), purge() 
  methods combination and to never let purge() be run with LOCK_log grabbed.
  
  
  
  split apart to allow
  the caller to chose either it
  Simulation of concurrently rotating/purging threads is added.
sql/log.h:
  new rotate(), purge() methods are added to be used instead of
  the former rotate_and_purge(RP_LOCK_LOG_IS_ALREADY_LOCKED).
  rotate_and_purge() signature is changed. Caller should not call rotate_and_purge()
  but rather {rotate(), purge()} if LOCK_log is acquired by it.
sql/rpl_injector.cc:
  changes to reflect the new rotate_and_purge() signature.
sql/sql_class.h:
  unnecessary constants are removed.
sql/sql_parse.cc:
  changes to reflect the new rotate_and_purge() signature.
sql/sql_reload.cc:
  changes to reflect the new rotate_and_purge() signature.
sql/sql_repl.cc:
  followup for bug@11747416: the file name printing is changed to a relative 
  path instead of just the file name.
This commit is contained in:
Andrei Elkin
2011-10-27 17:14:41 +03:00
parent bf456a8772
commit 95fdeb89c2
15 changed files with 238 additions and 51 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
mysql-test/suite/rpl/r/rpl_cant_read_event_incident.result
View File

@ -11,7 +11,7 @@ reset slave;
start slave;
include/wait_for_slave_param.inc [Last_IO_Errno]
Last_IO_Errno = '1236'
Last_IO_Error = 'Got fatal error 1236 from master when reading data from binary log: 'binlog truncated in the middle of event; consider out of disk space on master; the last event was read from 'master-bin.000001' at 316, the last byte read was read from 'master-bin.000001' at 335.''
Last_IO_Error = 'Got fatal error 1236 from master when reading data from binary log: 'binlog truncated in the middle of event; consider out of disk space on master; the last event was read from './master-bin.000001' at 316, the last byte read was read from './master-bin.000001' at 335.''
reset master;
stop slave;
reset slave;

2
mysql-test/suite/rpl/r/rpl_log_pos.result
View File

@ -9,7 +9,7 @@ change master to master_log_pos=MASTER_LOG_POS;
Read_Master_Log_Pos = '75'
start slave;
include/wait_for_slave_io_error.inc [errno=1236]
Last_IO_Error = 'Got fatal error 1236 from master when reading data from binary log: 'log event entry exceeded max_allowed_packet; Increase max_allowed_packet on master; the last event was read from 'master-bin.000001' at 75, the last byte read was read from 'master-bin.000001' at 94.''
Last_IO_Error = 'Got fatal error 1236 from master when reading data from binary log: 'log event entry exceeded max_allowed_packet; Increase max_allowed_packet on master; the last event was read from './master-bin.000001' at 75, the last byte read was read from './master-bin.000001' at 94.''
include/stop_slave_sql.inc
show master status;
File Position Binlog_Do_DB Binlog_Ignore_DB

2
mysql-test/suite/rpl/r/rpl_manual_change_index_file.result
View File

@ -5,7 +5,7 @@ CREATE TABLE t1(c1 INT);
FLUSH LOGS;
call mtr.add_suppression('Got fatal error 1236 from master when reading data from binary log: .*could not find next log');
include/wait_for_slave_io_error.inc [errno=1236]
Last_IO_Error = 'Got fatal error 1236 from master when reading data from binary log: 'could not find next log; the last event was read from 'master-bin.000002' at 237, the last byte read was read from 'master-bin.000002' at 237.''
Last_IO_Error = 'Got fatal error 1236 from master when reading data from binary log: 'could not find next log; the last event was read from './master-bin.000002' at 237, the last byte read was read from './master-bin.000002' at 237.''
CREATE TABLE t2(c1 INT);
FLUSH LOGS;
CREATE TABLE t3(c1 INT);

2
mysql-test/suite/rpl/r/rpl_packet.result
View File

@ -37,7 +37,7 @@ DROP TABLE t1;
CREATE TABLE t1 (f1 int PRIMARY KEY, f2 LONGTEXT, f3 LONGTEXT) ENGINE=MyISAM;
INSERT INTO t1(f1, f2, f3) VALUES(1, REPEAT('a', @@global.max_allowed_packet), REPEAT('b', @@global.max_allowed_packet));
include/wait_for_slave_io_error.inc [errno=1236]
Last_IO_Error = 'Got fatal error 1236 from master when reading data from binary log: 'log event entry exceeded max_allowed_packet; Increase max_allowed_packet on master; the last event was read from 'master-bin.000001' at 463, the last byte read was read from 'master-bin.000001' at 482.''
Last_IO_Error = 'Got fatal error 1236 from master when reading data from binary log: 'log event entry exceeded max_allowed_packet; Increase max_allowed_packet on master; the last event was read from './master-bin.000001' at 463, the last byte read was read from './master-bin.000001' at 482.''
STOP SLAVE;
RESET SLAVE;
RESET MASTER;

30
mysql-test/suite/rpl/r/rpl_rotate_purge_deadlock.result Normal file
View File

@ -0,0 +1,30 @@
include/master-slave.inc
[connection master]
show binary logs;
Log_name File_size
master-bin.000001 #
create table t1 (f text) engine=innodb;
SET DEBUG_SYNC = 'at_purge_logs_before_date WAIT_FOR rotated';
insert into t1 set f=repeat('a', 4096);
*** there must be two logs in the list ***
show binary logs;
Log_name File_size
master-bin.000001 #
master-bin.000002 #
insert into t1 set f=repeat('b', 4096);
*** there must be three logs in the list ***
show binary logs;
Log_name File_size
master-bin.000001 #
master-bin.000002 #
master-bin.000003 #
SET DEBUG_SYNC = 'now SIGNAL rotated';
SET DEBUG_SYNC = 'RESET';
SET DEBUG_SYNC = 'RESET';
SET DEBUG_SYNC = 'at_purge_logs_before_date WAIT_FOR rotated';
insert into t1 set f=repeat('b', 4096);
SET DEBUG_SYNC = 'now SIGNAL rotated';
SET DEBUG_SYNC = 'RESET';
SET DEBUG_SYNC = 'RESET';
drop table t1;
include/rpl_end.inc

5
mysql-test/suite/rpl/t/rpl_cant_read_event_incident.test
View File

@ -14,6 +14,11 @@
--source include/master-slave.inc
--source include/have_binlog_format_mixed.inc
#
# Bug#13050593 swallows `\' from Last_IO_Error
# todo: uncomment the filter once the bug is fixed.
#
--source include/not_windows.inc
call mtr.add_suppression("Error in Log_event::read_log_event()");

1
mysql-test/suite/rpl/t/rpl_rotate_purge_deadlock-master.opt Normal file
View File

@ -0,0 +1 @@
--max-binlog-size=4k --expire-logs-days=1

92
mysql-test/suite/rpl/t/rpl_rotate_purge_deadlock.test Normal file
View File

@ -0,0 +1,92 @@
#
# Bug#11763573 - 56299: MUTEX DEADLOCK WITH COM_BINLOG_DUMP, BINLOG PURGE, AND PROCESSLIST/KILL
#
source include/master-slave.inc;
source include/have_debug_sync.inc;
source include/have_binlog_format_row.inc;
source include/have_innodb.inc;
#
# Testing that execution of two concurrent INSERTing connections both
# triggering the binlog rotation is correct even though their execution
# is interleaved.
# The test makes the first connection to complete the rotation part
# and yields control to the second connection that rotates as well and
# gets first on purging. And the fact of interleaving does not create
# any issue.
#
connection master;
source include/show_binary_logs.inc;
create table t1 (f text) engine=innodb;
SET DEBUG_SYNC = 'at_purge_logs_before_date WAIT_FOR rotated';
send insert into t1 set f=repeat('a', 4096);
connection master1;
let $wait_condition=
SELECT COUNT(*) = 1 FROM INFORMATION_SCHEMA.PROCESSLIST
WHERE STATE like "debug sync point: at_purge_logs_before_date%";
--source include/wait_condition.inc
--echo *** there must be two logs in the list ***
source include/show_binary_logs.inc;
insert into t1 set f=repeat('b', 4096);
--echo *** there must be three logs in the list ***
source include/show_binary_logs.inc;
SET DEBUG_SYNC = 'now SIGNAL rotated';
SET DEBUG_SYNC = 'RESET';
# the first connection finally completes its INSERT
connection master;
reap;
SET DEBUG_SYNC = 'RESET';
sync_slave_with_master;
#
# Testing the reported deadlock involving DUMP, KILL and INSERT threads
#
connection master;
SET DEBUG_SYNC = 'at_purge_logs_before_date WAIT_FOR rotated';
send insert into t1 set f=repeat('b', 4096);
connection master1;
# make sure INSERT reaches waiting point
let $wait_condition=
SELECT COUNT(*) = 1 FROM INFORMATION_SCHEMA.PROCESSLIST
WHERE STATE like "debug sync point: at_purge_logs_before_date%";
--source include/wait_condition.inc
# find and kill DUMP thread
let $_tid= `select id from information_schema.processlist where command = 'Binlog Dump' limit 1`;
--disable_query_log
eval kill query $_tid;
--enable_query_log
#
# Now the proof is that the new DUMP thread has executed
# a critical section of the deadlock without any regression and is UP
#
let $wait_condition=
SELECT COUNT(*) = 1 FROM INFORMATION_SCHEMA.PROCESSLIST
WHERE command = 'Binlog Dump' and STATE like "Master has sent all binlog to slave%";
--source include/wait_condition.inc
SET DEBUG_SYNC = 'now SIGNAL rotated';
SET DEBUG_SYNC = 'RESET';
connection master;
reap;
SET DEBUG_SYNC = 'RESET';
drop table t1;
sync_slave_with_master;
--source include/rpl_end.inc