Bug #36979 merge-sync.test fails in pushbuild
Submitted: 26 May 2008 15:15 Modified: 19 Jul 2008 16:19
Reporter: Ingo Strüwing Email Updates:
Status: Closed Impact on me:
None 
Category:Tests: Engine Severity:S3 (Non-critical)
Version:6.0 OS:Any
Assigned to: Ingo Strüwing CPU Architecture:Any
Triage: D3 (Medium)

[26 May 2008 15:15] Ingo Strüwing
Description:
Seen on debx86-b in 6.0-engines-merge:

main.merge-sync                [ fail ]

--- /data0/pushbuild/pb1/pb/mysql-6.0-engines-merge/1/mysql-6.0.6-alpha-pb1/mysql-test/r/merge-sync.result	2008-05-23 11:53:13.000000000 +0300
+++ /data0/pushbuild/pb1/pb/mysql-6.0-engines-merge/1/mysql-6.0.6-alpha-pb1/mysql-test/r/merge-sync.reject	2008-05-23 13:09:56.596264809 +0300
@@ -114,7 +114,6 @@
 connection default;
 SELECT * FROM m1;
 c1
-2
 SET DEBUG_SYNC= 'RESET';
 DROP TABLE m1, t1;
 CREATE TABLE t1 (c1 INT) ENGINE= MyISAM;

mysqltest: Result content mismatch

I will disable the test case for now. Please re-enable after fix.

How to repeat:
See pushbuild.
[27 May 2008 19:49] 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/47108

ChangeSet@1.2644, 2008-05-27 21:49:04+02:00, istruewing@stella.local +4 -0
  Bug#36979 - merge-sync.test fails in pushbuild
  
  The test case failed randomly. When concurrent inserts are enabled,
  a SELECT in one connection does not always see rows inserted
  immediately before in another connection.
  See Bug#36618 - myisam insert not immediately visible to select
                  from another client.
  
  Fixed by disabling concurrent inserts for this test case.
[9 Jun 2008 16:33] Chuck Bell
Patch approved. Only 1 minor suggestion made for test. See commit email reply.
[19 Jul 2008 16:19] Ingo Strüwing
The core of the patch has been applied to the backup tree
during its merge to main. Only some comments proposed in this
patch aren't present in main. Since these are minor comments,
and the test case changed gravely due to meta data locking
changes, it doesn't seem worth the effort to modify the
comments so that they make sense in the current test case.
Hence closing the bug.