List:Commits« Previous MessageNext Message »
From:Andrei Elkin Date:March 8 2011 7:28pm
Subject:bzr commit into mysql-trunk branch (andrei.elkin:3737) Bug#58754 Bug#11765758
View as plain text  
#At file:///home/andrei/MySQL/BZR/2a-23May/mysql-trunk/ based on revid:build@stripped

 3737 Andrei Elkin	2011-03-08
      Bug#11765758  bug#58754 RPL_CORRUPTION FAILED on mysql-trunk
      
      There are two outstanding issues: a crash and stopping slave with an unexpected error.
      
      The current patch is hunting over the crash. Here is the first step to sort out
      whether `+d' dbug instruction that leads to the error-log relates to the crash.
      Notice, no real changes neither to the sources not the tests are done.
      The idea is just to watch PB2 on the win 64.
     @ mysql-test/suite/rpl/r/rpl_corruption.result
        results are updated.
     @ mysql-test/suite/rpl/t/rpl_corruption.test
        changed dbug control to use the abs not incremental syntax in order to avoid
        cluttering the error log with extra info.
        todo: restore to the original once relateness of the syntax to the issue got sorted out.

    modified:
      mysql-test/suite/rpl/r/rpl_corruption.result
      mysql-test/suite/rpl/t/rpl_corruption.test
=== modified file 'mysql-test/suite/rpl/r/rpl_corruption.result'
--- a/mysql-test/suite/rpl/r/rpl_corruption.result	2011-01-10 06:17:31 +0000
+++ b/mysql-test/suite/rpl/r/rpl_corruption.result	2011-03-08 19:28:55 +0000
@@ -33,10 +33,10 @@ START SLAVE IO_THREAD;
 include/wait_for_slave_io_error.inc [errno=1595]
 SET GLOBAL debug="-d,corrupt_queue_event";
 # 6. Slave. Corruption in relay log
-SET GLOBAL debug="+d,corrupt_read_log_event_char";
+SET GLOBAL debug="d,corrupt_read_log_event_char";
 START SLAVE;
 include/wait_for_slave_sql_error.inc [errno=1593]
-SET GLOBAL debug="-d,corrupt_read_log_event_char";
+SET GLOBAL debug="";
 # 7. Seek diff for tables on master and slave
 include/start_slave.inc
 include/diff_tables.inc [master:t1, slave:t1]

=== modified file 'mysql-test/suite/rpl/t/rpl_corruption.test'
--- a/mysql-test/suite/rpl/t/rpl_corruption.test	2011-01-10 06:17:31 +0000
+++ b/mysql-test/suite/rpl/t/rpl_corruption.test	2011-03-08 19:28:55 +0000
@@ -100,11 +100,19 @@ SET GLOBAL debug="-d,corrupt_queue_event
 
 # Emulate corruption in relay log
 --echo # 6. Slave. Corruption in relay log
-SET GLOBAL debug="+d,corrupt_read_log_event_char";
+
+#
+# Bug #58630 shows `+d' syntax has a side effect
+# Andrei changed it to unsigned `d' version which does not affect the tests logics.
+# todo: restore back to `+d' when  Bug #58630,  OBug#11765758 sorted out
+#
+SET GLOBAL debug="d,corrupt_read_log_event_char";
 START SLAVE;
 let $slave_sql_errno= 1593;
 --source include/wait_for_slave_sql_error.inc
-SET GLOBAL debug="-d,corrupt_read_log_event_char";
+# see above comments on `+d' syntax
+
+SET GLOBAL debug="";
 
 # Start normal replication and compare same table on master
 # and slave


Attachment: [text/bzr-bundle] bzr/andrei.elkin@oracle.com-20110308192855-glfphtgci36fg413.bundle
Thread
bzr commit into mysql-trunk branch (andrei.elkin:3737) Bug#58754 Bug#11765758Andrei Elkin8 Mar