[#TUC-130] In row replication, strings (varchar, text, ...) are stored in the mysql binlog using the table character set. If this character set differs from the JVM default character set, data can be corrupted at extraction time.

Issue Details (XML | Word)

Key:	TUC-130
Type:	Bug
Status:	Open
Priority:	Major
Assignee:	Stephane Giron
Reporter:	Stephane Giron
Votes:	0
Watchers:	0

Operations

If you were logged in you would be able to see more operations.

Tungsten Clustering

In row replication, strings (varchar, text, ...) are stored in the mysql binlog using the table character set. If this character set differs from the JVM default character set, data can be corrupted at extraction time.

Created: 29/Jul/10 04:36 AM Updated: 29/Jul/10 12:55 PM

Component/s:

DBMS - MySQL

Affects Version/s:

Release 1.2.3

Fix Version/s:

Release 1.3, Release 1.2.5

Original Estimate:

Unknown

Remaining Estimate:

Unknown

Time Spent:

Unknown

Sub-Tasks:

All

Open

Sub-Task Progress:

1.	Provide an option (in replicator.properties) for enabling / disabling this feature	Sub-task	Open	Stephane Giron
2.	Add a charset option in THL tool so that 'LIST' command can decode byte arrays into strings	Sub-task	Open	Stephane Giron

Description

Because of character set translation, if the character sets of the binlog and of the jvm do not match (for example latin1 vs. utf8), it is possible that data gets corrupted.
The idea is then to extract and keep in THL uninterpreted bytes instead of a java string.

If this data needs to be read, it could be done by using a filter.

All

Comments

Work Log

Change History

Sort Order:

There are no comments yet on this issue.