bzr commit into mysql-next-mr-bugfixing branch (Li-Bing.Song:3343) Bug#47699

[Li-Bing.Song]

#At file:///home/anders/Work/bzrwork/wt1/mysql-next-mr-bugfixing/ based on revid:horst.hunger@stripped

 3343 Li-Bing.Song@stripped	2010-11-10
      Bug#47699 rpl.rpl_backup_block fails sporadically
      
      It failed because an error appeared in Slave_SQL_Error,
      the test case expects that no error happens. The reason is
      that START SLAVE released the lock and returned before it cleared
      the error, so there was a possibility that Slave_SQL_Error was not
      0 when doing the following:
      START SLAVE SQL_THREAD;
      SHOW SLAVE STATUS;
      
      After this patch, clearing error always happens before START SLAVE
      releases the lock.

    modified:
      mysql-test/suite/rpl/r/rpl_slave_status.result
      mysql-test/suite/rpl/t/rpl_slave_status.test
      sql/rpl_slave.cc
=== modified file 'mysql-test/suite/rpl/r/rpl_slave_status.result'
--- a/mysql-test/suite/rpl/r/rpl_slave_status.result	2008-07-10 16:09:39 +0000
+++ b/mysql-test/suite/rpl/r/rpl_slave_status.result	2010-11-10 05:58:14 +0000
@@ -29,7 +29,32 @@ include/stop_slave.inc
 START SLAVE;
 ==== Verify that Slave_IO_Running = No ====
 Slave_IO_Running = No (should be No)
-==== Cleanup (Note that slave IO thread is not running) ====
+==== Cleanup ====
+[on master]
+DROP TABLE t1;
+[on slave]
+include/stop_slave.inc
+CHANGE MASTER TO master_user='root', master_password='';
+include/start_slave.inc
+
+# Bug#47699 rpl.rpl_backup_block fails sporadically
+#
+# START SLAVE released the lock and returned before it cleared the error,
+# so there is a possibility that Slave_SQL_Error is not 0.
+[on slave]
+CALL mtr.add_suppression("Slave: Table 't1' already exists Error_code: 1050");
+# The statement makes SQL thread to fail.
+CREATE TABLE t1(c1 INT);
+[on master]
+CREATE TABLE t1(c1 INT);
+[on slave]
 DROP TABLE t1;
+# Block SQL thread immediately after it starts.
+SET GLOBAL debug="d,after_start_slave";
+START SLAVE SQL_THREAD;
+# Check Slave_SQL_Error, there should not be an error.
+# Resume SQL thread
+SET DEBUG_SYNC="now SIGNAL signal.continue";
+SET GLOBAL debug= $debug_save;
 [on master]
 DROP TABLE t1;

=== modified file 'mysql-test/suite/rpl/t/rpl_slave_status.test'
--- a/mysql-test/suite/rpl/t/rpl_slave_status.test	2009-10-02 09:24:21 +0000
+++ b/mysql-test/suite/rpl/t/rpl_slave_status.test	2010-11-10 05:58:14 +0000
@@ -60,11 +60,53 @@ source include/wait_for_slave_io_to_stop
 let $result= query_get_value("SHOW SLAVE STATUS", Slave_IO_Running, 1);
 --echo Slave_IO_Running = $result (should be No)
 
---echo ==== Cleanup (Note that slave IO thread is not running) ====
+--echo ==== Cleanup ====
+--echo [on master]
+connection master;
+DROP TABLE t1;
+
+--echo [on slave]
+connection slave;
+source include/stop_slave.inc;
+CHANGE MASTER TO master_user='root', master_password='';
+source include/start_slave.inc;
+connection master;
+sync_slave_with_master;
+
+--echo 
+--echo # Bug#47699 rpl.rpl_backup_block fails sporadically
+--echo #
+--echo # START SLAVE released the lock and returned before it cleared the error,
+--echo # so there is a possibility that Slave_SQL_Error is not 0.
+
+--echo [on slave]
+CALL mtr.add_suppression("Slave: Table 't1' already exists Error_code: 1050");
+--echo # The statement makes SQL thread to fail.
+CREATE TABLE t1(c1 INT);
+
+--echo [on master]
+connection master;
+CREATE TABLE t1(c1 INT);
+
+--echo [on slave]
+connection slave;
+source include/wait_for_slave_sql_to_stop.inc;
+
 DROP TABLE t1;
-# cleanup: slave io thread has been stopped "irrecoverably"
-# so we clean up mess manually
+
+--echo # Block SQL thread immediately after it starts.
+let $debug_save=`SELECT @@debug`;
+SET GLOBAL debug="d,after_start_slave";
+START SLAVE SQL_THREAD;
+source include/wait_for_slave_sql_to_start.inc;
+--echo # Check Slave_SQL_Error, there should not be an error.
+source include/check_slave_no_error.inc;
+
+--echo # Resume SQL thread
+SET DEBUG_SYNC="now SIGNAL signal.continue";
+SET GLOBAL debug= $debug_save;
 
 --echo [on master]
 connection master;
 DROP TABLE t1;
+source include/master-slave-end.inc;

=== modified file 'sql/rpl_slave.cc'
--- a/sql/rpl_slave.cc	2010-10-27 10:31:46 +0000
+++ b/sql/rpl_slave.cc	2010-11-10 05:58:14 +0000
@@ -3465,8 +3465,6 @@ pthread_handler_t handle_slave_sql(void
     Seconds_Behind_Master grows. No big deal.
   */
   rli->abort_slave = 0;
-  mysql_mutex_unlock(&rli->run_lock);
-  mysql_cond_broadcast(&rli->start_cond);
 
   /*
     Reset errors for a clean start (otherwise, if the master is idle, the SQL
@@ -3480,6 +3478,19 @@ pthread_handler_t handle_slave_sql(void
   */
   rli->clear_error();
 
+  mysql_mutex_unlock(&rli->run_lock);
+  mysql_cond_broadcast(&rli->start_cond);
+
+  DBUG_EXECUTE_IF("after_start_slave",
+                  {
+                    const char act[]=
+                      "now "
+                      "wait_for signal.continue";
+                    DBUG_ASSERT(opt_debug_sync_timeout > 0);
+                    DBUG_ASSERT(!debug_sync_set_action(current_thd,
+                                                       STRING_WITH_LEN(act)));
+                  };);
+
   //tell the I/O thread to take relay_log_space_limit into account from now on
   mysql_mutex_lock(&rli->log_space_lock);
   rli->ignore_log_space_limit= 0;


Attachment: [text/bzr-bundle] bzr/[e-mail]