Bug #32720 Test rpl_ndb_mix_innodb fails on SPARC and PowerPC
Submitted: 26 Nov 2007 14:29 Modified: 19 Jun 2009 12:24
Reporter: Mats Kindahl Email Updates:
Status: No Feedback Impact on me:
None 
Category:MySQL Cluster: Replication Severity:S2 (Serious)
Version:mysql-5.1 OS:Solaris (SPARC and PowerPC only)
Assigned to: Assigned Account CPU Architecture:Any
Tags: 5.1.23-bk

[26 Nov 2007 14:29] Mats Kindahl
Description:
Test case rpl_ndb_mix_innodb fails on Solaris 10 SPARC machine and the PowerPC OS X machine. The failure manifests itself by a DATETIME column being garbage. Here's an example:

  INSERT INTO `history`
  VALUES (1,74,10,10,'1.00','�(+(-,0-)0 83:90:40', 'root@localhost',
          '868eddd8-ebff-102a-b111-b4af4f80966c', 'completed trans');

where the table has the definition

  CREATE TABLE tpcb.history (id MEDIUMINT NOT NULL AUTO_INCREMENT,aid INT,
                             tid INT, bid INT,  amount DECIMAL(10,2),
                             tdate DATETIME, teller CHAR(20), uuidf LONGBLOB,
                             filler CHAR(80),PRIMARY KEY (id));

The test case replicates from an InnoDB table to an NDB table, and only the DATETIME column differs. The replication is done using row-based replication, because the original statement in the test holds a call to NOW() in the DATETIME column.

When trying to trigger the failure manually by executing the correct insert on the master, replication works without problems.

How to repeat:
Look at rpl_ndb_mix_innodb on either sol10-sparc-a or powermac5.

Suggested fix:
Make sure that dates are replicated correctly.
[19 May 2009 12:24] Jonathan Miller
Is this still a valid bug? Please close if not.
[19 Jun 2009 23:00] Bugs System
No feedback was provided for this bug for over a month, so it is
being suspended automatically. If you are able to provide the
information that was originally requested, please do so and change
the status of the bug back to "Open".