MySQL Bugs: #14872: Random test failure in the skip_name

Bug #14872	Random test failure in the skip_name_resolve test
Submitted:	11 Nov 2005 17:30	Modified:	12 Dec 2005 13:48
Reporter:	Lenz Grimmer	Email Updates:
Status:	Closed	Impact on me:	None
Category:	MySQL Server	Severity:	S3 (Non-critical)
Version:	5.0.16-pre	OS:	Linux (Linux)
Assigned to:	Kent Boortz	CPU Architecture:	Any

Description:
The "skip_name_resolve" test suite fails every once in a while when using the Perl version of mysql-test-run with a diff similar to that one:

TEST                           RESULT
-------------------------------------------------------

skip_name_resolve              [ fail ]

Errors are (from /home/mysqldev/bug14822/mysql-pro-gpl-5.0.16-linux-i686-csc7204/mysql-test/var/log/mysqltest-time) :
mysqltest: Result length mismatch
(the last lines may be the most important ones)
Below are the diffs between actual and expected results:
-------------------------------------------------------
*** r/skip_name_resolve.result  2005-11-11 14:50:34.000000000 +0300
--- r/skip_name_resolve.reject  2005-11-11 20:03:22.000000000 +0300
***************
*** 10,14 ****
  #
  show processlist;
  Id    User    Host    db      Command Time    State   Info
! #     root    #       test    Sleep   #               NULL
  #     root    #       test    Query   #       NULL    show processlist
--- 10,14 ----
  #
  show processlist;
  Id    User    Host    db      Command Time    State   Info
! #     root    #       test    Query   #       Writing to net  DROP USER mysqltest_1@'127.0.0.1/255.255.255.255'
  #     root    #       test    Query   #       NULL    show processlist
-------------------------------------------------------

The Shell version does not reveal this diff.

How to repeat:
Run "mysql-test-run.pl skip_name_resolve" several times. Observe the diff. It seems to be a timing issue that only surfaces when using the Perl mysql-test-run script.

This bug has been fixed in mysql-5.0 by this changeset:

    ChangeSet 1.1960.79.1 2005/12/07 17:01:17 anozdrin@mysql.com

The fix masks out some columns that are not predictable in the test.

The fix is not in 5.0.17, but will be in 5.0.18.

Extracted the part from the larger 5.0 patch that solved this bug,
and applied to 5.0.17. So it will be in the 5.0.17 release.