MySQL Bugs: #8585: ndb_cache2 fails on aix52

Bug #8585	ndb_cache2 fails on aix52
Submitted:	17 Feb 2005 17:48	Modified:	11 Apr 2005 2:00
Reporter:	Tomas Ulin	Email Updates:
Status:	Closed	Impact on me:	None
Category:	MySQL Cluster: Cluster (NDB) storage engine	Severity:	S2 (Serious)
Version:	5.0.4	OS:	Linux (Linux + Solaris)
Assigned to:	Magnus Blåudd	CPU Architecture:	Any

Description:
run mysql-test-run

see also internal info

How to repeat:
run mysql-test-run

Pushed to mysql-5.0

Issue still occurs in build based on ChangeSet
  1.1806 05/03/13 01:07:01 igor@rurik.mysql.com +1 -0
  join_outer.test:
    Correction after manual merge.

AIX52 did not build in that snapshot, but this issue occurs on
- ita2 (ia64, Linux) in default mode only, PS test passes;
- sunfire100b (Solaris 9, both 32 and 64 bit) in PS mode only, default test passes;
- sunfire100c (Solaris 8) in 64 bit only, both test modes; 32 bit passes.

Tree saved on ita2 in directory bug8585.

I have now pushed a fix for this that makes ndb_cache2 NOT timing dependent, this should fix this problem.

https://intranet.mysql.com/~kboortz/status/mysql-5.0.3-alpha/logs/bsd53-5.0-max.log.html>>
10762: ndb_cache2                     [ fail ]
10763: 
10764: Errors are (from /usr/home/mysqldev/bsd53/test/mysql-max-5.0.3-alpha-unknown-freebsd5.3-i386/mysql-test/var/log/mysqltest-time) :
10765: /usr/home/mysqldev/bsd53/test/mysql-max-5.0.3-alpha-unknown-freebsd5.3-i386/bin/mysqltest: At line 361: Result content mismatch
10766: (the last lines may be the most important ones)
10767: Below are the diffs between actual and expected results:
10768: -------------------------------------------------------
10769: *** r/ndb_cache2.result	Mon Mar 21 01:35:31 2005
10770: --- r/ndb_cache2.reject	Mon Mar 21 18:08:23 2005
10771: ***************
10772: *** 122,128 ****
10773:   1	3	3	First row
10774:   show status like "Qcache_hits";
10775:   Variable_name	Value
10776: ! Qcache_hits	4
10777:   update t1 set a1=4 where b1=3;
10778:   use test;
10779:   show status like "Qcache_queries_in_cache";
10780: --- 122,128 ----
10781:   1	3	3	First row
10782:   show status like "Qcache_hits";
10783:   Variable_name	Value
10784: ! Qcache_hits	3
10785:   update t1 set a1=4 where b1=3;
10786:   use test;
10787:   show status like "Qcache_queries_in_cache";
10788: ***************
10789: *** 138,147 ****
10790:   1	4	3	First row
10791:   show status like "Qcache_inserts";
10792:   Variable_name	Value
10793: ! Qcache_inserts	7
10794:   show status like "Qcache_hits";
10795:   Variable_name	Value
10796: ! Qcache_hits	5
10797:   select * from t1 order by pk desc;
10798:   pk	a1	b1	c1
10799:   2	7	8	Second row
10800: --- 138,147 ----
10801:   1	4	3	First row
10802:   show status like "Qcache_inserts";
10803:   Variable_name	Value
10804: ! Qcache_inserts	8
10805:   show status like "Qcache_hits";
10806:   Variable_name	Value
10807: ! Qcache_hits	4
10808:   select * from t1 order by pk desc;
10809:   pk	a1	b1	c1
10810:   2	7	8	Second row
10811: ***************
10812: *** 155,164 ****
10813:   Qcache_queries_in_cache	1
10814:   show status like "Qcache_inserts";
10815:   Variable_name	Value
10816: ! Qcache_inserts	7
10817:   show status like "Qcache_hits";
10818:   Variable_name	Value
10819: ! Qcache_hits	7
10820:   select * from t2;
10821:   pk	a2	b2
10822:   1	2	3
10823: --- 155,164 ----
10824:   Qcache_queries_in_cache	1
10825:   show status like "Qcache_inserts";
10826:   Variable_name	Value
10827: ! Qcache_inserts	8
10828:   show status like "Qcache_hits";
10829:   Variable_name	Value
10830: ! Qcache_hits	6
10831:   select * from t2;
10832:   pk	a2	b2
10833:   1	2	3
10834: -------------------------------------------------------
10835: Please follow the instructions outlined at
10836: http://www.mysql.com/doc/en/Reporting_mysqltest_bugs.html
10837: to find the reason to this problem and how to report this.
10838: 
10839: 
10840: Ending Tests
10841: Shutting-down MySQL daemon

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/internals/23526

Commit count cache was not properly invalidated when deleting a record using cursor.

A bug fix for the problem has been pushed to 5.0.4.

Noted in 5.0.4 changelog.