MySQL Bugs: #21364: IM test failure: "The instance is already started" (in "im_life

Bug #21364	IM test failure: "The instance is already started" (in "im_life_cycle")
Submitted:	31 Jul 2006 13:29	Modified:	31 Aug 2006 8:51
Reporter:	Joerg Bruehe	Email Updates:
Status:	Can't repeat	Impact on me:	None
Category:	Instance Manager	Severity:	S3 (Non-critical)
Version:	5.0.24	OS:	Other (OpenBSD)
Assigned to:	Alexander Nozdrin	CPU Architecture:	Any

Description:
Occurred during the release build of 5.0.24.

This sure looks like bug#14106, but both patches for that bug are contained in the sources for building (double-checked)!

This bug is specific to OpenBSD (the symptom occure nowhere else),
and all test runs there fail from this bug:

=====
im_life_cycle                  [ fail ]

Errors are (from /PATH/mysqltest-time) :
mysqltest: At line NNN: query 'START INSTANCE mysqld2' failed: 3002: The instance is already started
(the last lines may be the most important ones)
Result from queries before failure can be found in r/im_life_cycle.log

Killing Possible Leftover Processes
=====

granny-5.0-cluster.log   normal
granny-5.0-cluster.log   NDB
granny-5.0-max.log   normal
granny-5.0-max.log   NDB
granny-5.0-pro.log   normal
granny-5.0-standard.log   normal
granny-5.0-debug.log   normal
granny-5.0-pro-gpl.log   normal

The test seems to leave the server in some problematic state,
because all following "index_merge*" and "information_schema*" tests
fail with connect issues,
until test "init_connect" seems to cause a re-init.

Surprisingly, tests "im_options_set", "im_options_unset", and "im_utils"
pass in between.    

How to repeat:
Detected by running the test suite.

This is exposing a regression--thus a P1.

This bug seems to be fixed by the patch for BUG#21331:
http://lists.mysql.com/commits/10998

(fixed in 5.0.24a)