Details
-
Bug
-
Status: Open (View Workflow)
-
Major
-
Resolution: Unresolved
-
10.8.1, 10.9.1, 10.10.1, 10.11.1, 11.0.1
Description
If the (parallel) slave fails while applying START ALTER, a memory leak of
the start_alter_info results. This is because the start_alter_info object is
added into the mi->start_alter_list in write_bin_log_start_alter(), however
if the START ALTER fails, this is not called.
I will push a fix for this on my branch knielsen_start_alter on github.
Test case:
--source include/have_innodb.inc
|
--source include/master-slave.inc
|
|
|
--connection master
|
CREATE TABLE t1 (a INT PRIMARY KEY);
|
|
|
--sync_slave_with_master
|
--source include/stop_slave.inc
|
DROP TABLE t1;
|
SET GLOBAL slave_parallel_threads = 5;
|
--source include/start_slave.inc
|
|
|
--connection master
|
SET binlog_alter_two_phase = ON;
|
ALTER TABLE t1 ADD COLUMN b INT DEFAULT NULL;
|
|
|
--connection slave
|
# 1146 = ER_NO_SUCH_TABLE
|
--let $slave_sql_errno= 1146
|
--source include/wait_for_slave_sql_error.inc
|
CREATE TABLE t1 (a INT PRIMARY KEY);
|
--source include/start_slave.inc
|
|
|
--connection master
|
DROP TABLE t1;
|
--source include/rpl_end.inc
|