From: Date: March 8 2008 11:15am
Subject: bk commit into 6.0 tree (aelkin:1.2590) BUG#34686
List-Archive: http://lists.mysql.com/commits/43642
X-Bug: 34686
Message-Id: <200803081015.m28AFr7R012642@mysql1000>

Below is the list of changes that have just been committed into a local
6.0 repository of aelkin.  When aelkin does a push these changes
will be propagated to the main repository and, within 24 hours after the
push, to the public repository.
For information on how to access the public repository
see http://dev.mysql.com/doc/mysql/en/installing-source-tree.html

ChangeSet@stripped, 2008-03-08 12:14:47+02:00, aelkin@mysql1000.(none) +2 -0
  Bug #34686 change master w/o explict master_heartbeat_period dont reset
                 period to default
      
      The original patch has been pushed to 5.1-telco-6.3.
      This separate one is for 6.0 because merging is in reverse:
      6.0->telco.
      
  CHANGE Master w/o the explicit  master_heartbeat_period is supposed to
  reset the period to the value of @@global.slave_net_timeout/2.
  However, there was a possibility to get the hb status un-reset preserving its
  former value.
  Fixed with adding a rule to reset to the default in change_master().
  sql/sql_repl.cc
  ensuring hb repiod gets reset to the specified default.

  mysql-test/suite/rpl/t/rpl_heartbeat.test@stripped, 2008-03-08 12:14:45+02:00, aelkin@mysql1000.(none) +10 -1
    adding explanation to disable_warnings instuction.

  sql/sql_repl.cc@stripped, 2008-03-08 12:14:45+02:00, aelkin@mysql1000.(none) +3 -0
    ensuring hb repiod gets reset to the specified default.

diff -Nrup a/mysql-test/suite/rpl/t/rpl_heartbeat.test b/mysql-test/suite/rpl/t/rpl_heartbeat.test
--- a/mysql-test/suite/rpl/t/rpl_heartbeat.test	2008-02-03 11:00:47 +02:00
+++ b/mysql-test/suite/rpl/t/rpl_heartbeat.test	2008-03-08 12:14:45 +02:00
@@ -16,8 +16,17 @@ connection master;
 reset master;
 
 connection slave;
-set @@global.slave_net_timeout= 10;
 
+# there can be a warning if mtr set the global init value bigger than
+# the following assignment. The warning is a grace for the user
+# to warn her that the new value of the global var is bigger than
+# the current value of the hb period, i.e the sanity condition
+# 0 < hb < slave_net_timeout does not hold.
+# The warning has to be disabled as mtr can change the init value
+# for slave_net_timeout
+--disable_warnings
+set @@global.slave_net_timeout= 10;
+--enable_warnings
 ###
 ### Checking the range
 ###
diff -Nrup a/sql/sql_repl.cc b/sql/sql_repl.cc
--- a/sql/sql_repl.cc	2008-02-19 14:59:09 +02:00
+++ b/sql/sql_repl.cc	2008-03-08 12:14:45 +02:00
@@ -1255,6 +1255,9 @@ bool change_master(THD* thd, Master_info
     mi->connect_retry = lex_mi->connect_retry;
   if (lex_mi->heartbeat_opt != LEX_MASTER_INFO::LEX_MI_UNCHANGED)
     mi->heartbeat_period = lex_mi->heartbeat_period;
+  else
+    mi->heartbeat_period= (float) min(SLAVE_MAX_HEARTBEAT_PERIOD,
+                                      (slave_net_timeout/2.0));
   mi->received_heartbeats= LL(0); // counter lives until master is CHANGEd
   if (lex_mi->ssl != LEX_MASTER_INFO::LEX_MI_UNCHANGED)
     mi->ssl= (lex_mi->ssl == LEX_MASTER_INFO::LEX_MI_ENABLE);