Merge lp:~laurynas-biveinis/percona-server/bug1361568 into lp:percona-server/5.6

[Laurynas Biveinis]

Fix bug 1361568 (Intermittent rpl_err_ignoredtable failures on the 5.6
trunk).

The bug is in the testcase, the relevant bits are:

connection master1;
select get_lock('crash_lock%20C', 10);

connection master;
...
send update t2 set a = a + 1 + get_lock('crash_lock%20C', 10);

connection master1;
...
kill @id;
drop table t2,t3;
...
connection master;
# The get_lock function causes warning for unsafe statement.
--disable_warnings
--error 0,1317,2013
reap;

KILL and DROP TABLE statements race and may end up in the master1
thread holding the user lock and attempting to drop the tables, while
the master thread hasn't reacted to the kill bit yet, and thus has the
same tables locked and is waiting for the user lock.

Fix with a MariaDB patch that waits for the KILL to complete before
proceeding with the DROP TABLE
(https://github.com/MariaDB/server/commit/45f8bcb282f9ff68c2a21366aad5947380a503c0).

http://jenkins.percona.com/job/percona-server-5.6-param/700/

[Laurynas Biveinis (laurynas-biveinis)]

review: Approve

=== modified file 'mysql-test/suite/rpl/t/rpl_err_ignoredtable.test'
--- mysql-test/suite/rpl/t/rpl_err_ignoredtable.test	2014-06-11 13:33:46 +0000
+++ mysql-test/suite/rpl/t/rpl_err_ignoredtable.test	2014-08-27 06:15:43 +0000
@@ -47,6 +47,9 @@
 source include/wait_condition.inc;
 select (@id := id) - id from t3;
 kill @id;
+let $id= `SELECT @id`;
+let $wait_condition= SELECT COUNT(*) = 0 FROM INFORMATION_SCHEMA.PROCESSLIST WHERE ID = $id;
+source include/wait_condition.inc;
 drop table t2,t3;
 insert into t4 values (3),(4);
 connection master;