Bug #45132 Replication failure: 'Table doesn't exist' on slave

• Submitted: 27 May 2009 14:06
• Modified: 26 Nov 2009 16:47
• Reporter: Philip Stoev
• Status: Duplicate
• Category: MySQL Server: Replication
• Severity: S3 (Non-critical)
• Version: 6.0
• OS: Any
• Assigned to: Assigned Account
• CPU Architecture: Any

[27 May 2009 14:06] Philip Stoev

Description:
When executing a random workload, involving various statements of interest to replication, dynamic changes to the binlog format, and DDL, the slave failed as follows:

 Error 'Table 'test.x' doesn't exist' on query. Default database: 'test'. Query: 'UPDATE `x` SET `float_unsigned` = 2 WHERE `int_not_null_key` = 0 ORDER BY `bit_key` LIMIT 8'

Where X is a table created and possibly dropped during the test. This failure was only observed on 6.0/Azalea and not on 5.1

See also bug #40188 that exhibits the same failure, but appears to be fixed -- the test case for bug #40188 no longer fails with this failure.

How to repeat:
The binary log is availabe. A repeatable test case will be provided once the test case has been narrowed down sufficiently to become more repeatable.

[26 Nov 2009 16:47] Philip Stoev

See bug #47151 instead.