Bug #43647 Debug message "[Note] new_mode 0" floods slave error log

• Submitted: 14 Mar 2009 10:02
• Modified: 25 Jun 2009 11:00
• Reporter: Philip Stoev
• Status: Closed
• Category: MySQL Server: Replication
• Severity: S3 (Non-critical)
• Version: 6.0.10
• OS: Any
• Assigned to: Luis Soares
• CPU Architecture: Any

[14 Mar 2009 10:02] Philip Stoev

Description:
When executing a replication workload, the slave error log is flooded with messages like :

090309 10:25:32 [Note] new_mode 0

On a moderately loaded server, approximately 3600 such messages are printed in the log *per second*. On systems where log rotation of the error log is not in place, this will cause unexpected exhaustion of hard disk space.

How to repeat:
A test case will be uploaded shortly.

[14 Mar 2009 10:46] Philip Stoev

This message is printed for every binary log event by line 2793 of ./sql/slave.cc . I am not providing a test case because the issue is obvious from a code inspection. Setting directly to verified.

The sql_print_information() function was used to print this message. Instead, DBUG_PRINT() appears to be more appropriate.

[4 May 2009 9:22] Bugs System

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/73282

2849 Luis Soares	2009-05-04
      BUG#43647: Debug message "[Note] new_mode 0" floods slave error log
      
      There was a print out (through sql_print_warning) inside
      handle_slave_sql loop, right before the exec_relay_log_event. This
      would output several entries in mysqld error log, eventually causing
      cause unexpected exhaustion of disk space. Output line sample:
      
      "090309 10:25:32 [Note] new_mode 0".
      
      This patch fixes this by removing the call to sql_print_warning.

[13 May 2009 10:55] Bugs System

A patch for this bug has been committed. After review, it may
be pushed to the relevant source trees for release in the next
version. You can access the patch from:

  http://lists.mysql.com/commits/73929

2849 Luis Soares	2009-05-13
      BUG#43647: Debug message "[Note] new_mode 0" floods slave error log
      
      There was a print out (through sql_print_warning) inside
      handle_slave_sql loop, right before the exec_relay_log_event. This
      would output several entries in mysqld error log, eventually causing
      unexpected exhaustion of disk space. Output line sample:
            
      "090309 10:25:32 [Note] new_mode 0".
           
      This patch fixes this by removing the call to sql_print_warning.

[13 May 2009 10:55] Luis Soares

The second patch is just to a fix in commit message.

[13 May 2009 10:58] Luis Soares

Pushed to 6.0-rpl.

[16 Jun 2009 14:57] Bugs System

Pushed into 5.4.4-alpha (revid:[e-mail]-20090612081554-kmu72j3wenjn82a8) (version source revid:[e-mail]-20090612034751-dqd4y0bv0iln1yaj) (merge vers: 5.4.4-alpha) (pib:7)

[17 Jun 2009 19:28] Bugs System

Pushed into 5.4.4-alpha (revid:[e-mail]-20090616183122-chjzbaa30qopdra9) (version source revid:[e-mail]-20090612034751-dqd4y0bv0iln1yaj) (merge vers: 5.4.4-alpha) (pib:11)

[25 Jun 2009 11:00] Jon Stephens

Documented bugfix in the 5.4.4 changelog as follows:

        An extraneous debug message was written repeatedly to the slave
        error log, which could eventually lead to premature exhaustion
        of disk space.