[MDEV-734] LP:917662 - mysql_print_status() missing final fflush() Created: 2012-01-17  Updated: 2020-05-12  Resolved: 2014-03-22

Status: Closed
Project: MariaDB Server
Component/s: None
Affects Version/s: 5.1.67, 5.2.14, 5.3.12, 5.5.36, 10.0.9
Fix Version/s: 10.0.10

Type: Bug Priority: Minor
Reporter: Hartmut Holzgraefe Assignee: Michael Widenius
Resolution: Fixed Votes: 0
Labels: Launchpad, upstream

Attachments: XML File LPexportBug917662.xml    

 Description   

See http://bugs.mysql.com/bug.php?id=64048

Description:
When requesting debug status output with

mysqladmin debug

the output in the error log (assuming that mysqld_safe was used)
seems to be truncated at first as the mallinfo() related part
is missing, this part will only show up in the error log when
debug is called a 2nd time. This makes the missing part of the
first invocation appear but then again the mallinfo() output
for the 2nd call is missing

How to repeat:
Issue "mysqladmin debug", check output in mysql error log for the "Memory status:"
section ...

Suggested fix:
add an extra fflush(stdout) at the very end of mysql_print_status() in sql/sql_test.cc



 Comments   
Comment by Rasmus Johansson (Inactive) [ 2012-03-15 ]

Launchpad bug id: 917662

Comment by Elena Stepanova [ 2013-03-10 ]

Fixed in 5.6.5

Comment by Elena Stepanova [ 2014-03-18 ]

The upstream bug was fixed in 5.6 with the following revision:

    revno: 3402.50.249
    revision-id: praveenkumar.hulakund@oracle.com-20120224064323-7syqusja74saixbx
    parent: chaithra.gopalareddy@oracle.com-20120224063144-rmkfz8hak2oq1id4
    committer: Praveenkumar Hulakund <praveenkumar.hulakund@oracle.com>
    branch nick: trunk
    timestamp: Fri 2012-02-24 12:13:23 +0530
    message:
      Bug#13608112 - 64048: MYSQL_PRINT_STATUS() MISSING FINAL FFLUSH()
      
      Analysis:
      While printing debug information, memory related information and event
      related information were written to buffer after flushing it. Because of this,
      these information were getting printed when we do the next flush.
      
      Fix:
      Instead of calling fflush() before writing memory and event
      information, calling it after memory and event information.

I suppose it will be fixed by the current activity of merging bugfixes from 5.6.

Comment by Michael Widenius [ 2014-03-22 ]

Fixed in my 10.0 tree. Will be pushed shortly

Generated at Thu Feb 08 06:30:56 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.