As noted by knielsen, this test suffers sporadic failures. I was able to repeat against the latest mysql-5.3.
All the failures seem to have the following in common:
* the diff reports that the master has 2 or more rows more than the slave;
* the divering tables are always the ones that have no PK
As noted by knielsen, this test suffers sporadic failures. I was able to repeat against the latest mysql-5.3.
All the failures seem to have the following in common:
* the diff reports that the master has 2 or more rows more than the slave;
* the divering tables are always the ones that have no PK
Logs will be attached shortly.