MySQL Bugs: #31848: Test failure: Cluster has problems on insert with auto-increment

Bug #31848	Test failure: Cluster has problems on insert with auto-increment
Submitted:	25 Oct 2007 13:20	Modified:	5 Nov 2007 23:08
Reporter:	Joerg Bruehe	Email Updates:
Status:	Closed	Impact on me:	None
Category:	MySQL Cluster: Cluster (NDB) storage engine	Severity:	S3 (Non-critical)
Version:	5.0.44 (NDB) + x	OS:	Any (gcc 4.2.1, x86 (32))
Assigned to:	Kristian Nielsen	CPU Architecture:	Any

Description:
Found in the "mysql-5.0-ndb" tree
when checking the fix of bug#31761 (yes, error 879 is gone):

ndb_alter_table                [ fail ]

Errors are (from /MySQL/M50/ndb-5.0/mysql-test/var/log/mysqltest-time) :
mysqltest: At line 62: query 'insert into t1 values
(0,4,3,5,"PENDING",1,7),(NULL,4,3,5,"PENDING",1,7),(31,4,3,5,"PENDING",1,7), (7,4,3,5,"PENDING",1,7), (NULL,4,3,5,"PENDING",1,7), (100,4,3,5,"PENDING",1,7), (99,4,3,5,"PENDING",1,7), (8,4,3,5,"PENDING",1,7), (NULL,4,3,5,"PENDING",1,7)' failed: 1467: Failed to read auto-increment value from storage engine
(the last lines may be the most important ones)
Result from queries before failure can be found in r/ndb_alter_table.log

ndb_replace                    [ fail ]

Errors are (from /MySQL/M50/ndb-5.0/mysql-test/var/log/mysqltest-time) :
mysqltest: At line 57: query 'REPLACE INTO t2 (a,b,c,x,y,z,i) VALUES (1,1,1,'a','a','a',1),(1,1,1,'b','b','b',2), (1,1,1,'c','c','c',3)' failed: 1467: Failed to read auto-increment value from storage engine
(the last lines may be the most important ones)
Result from queries before failure can be found in r/ndb_replace.log

Latest changeset:
ChangeSet@1.2489, 2007-10-25 10:54:23+02:00, knielsen@loke.(none) +1 -0
  Merge loke.(none):/home/knielsen/devel/bug31810
  into  loke.(none):/home/knielsen/devel/mysql-5.0-ndb
  MERGE: 1.2486.1.1

Build was done using 'compile-pentium-max', this uses "-O3".

How to repeat:
Found by running the test suite.

I repeated this using "-O2" (same sources), and the bugs were gone.

I managed to repeat this on 32-bit Ubuntu feisty x86, by compiling gcc 4.2.1.

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/36820

ChangeSet@1.2490, 2007-11-01 07:59:55+01:00, knielsen@ymer.(none) +1 -0
  Bug #31848: Test failure: Cluster has problems on insert with auto-increment
  
  Fix uninitialized variable causing failures for some interpreted update
  operations on gcc 4.2.1.

This was introduced with the fix for BUG#27437. So the bug would have first appeared in 5.0.42.

It only applies to 5.0, the 5.1+ code does not have the problem.

Symptoms would be randomly failing interpreted update operations (this includes auto increments, which use interpreted updates internally).

Pushed to mysql-5.0-ndb (and null-merged to mysql-5.1-new-ndb).

Pushed into 6.0.4-alpha

Pushed into 5.1.23-rc

Pushed into 5.0.52

Thank you for your bug report. This issue has been committed to our source repository of that product and will be incorporated into the next release.

If necessary, you can access the source repository and build the latest available version, including the bug fix. More information about accessing the source trees is available at

    http://dev.mysql.com/doc/en/installing-source.html

Documented in 5.0.52 changelog (5.0.52 only - per developer note above). Closed.

Fix also documented in 5.1.22-ndb-6.28 and 5.1.22-ndb-6.3.6 changelogs (per release notes from Tomas).

Removed 5.1.22-ndb-6.2.8 and 5.1.22-ndb-6.3.6 changelog entries after KNielsen & Tomas confirmed null merge to 5.1/CGE trees.