Bug #33210 | SBR & read-committed / read-uncommitted transaction isolations yield an error | ||
---|---|---|---|
Submitted: | 13 Dec 2007 12:55 | Modified: | 20 Dec 2007 1:01 |
Reporter: | Domas Mituzas | Email Updates: | |
Status: | Closed | Impact on me: | |
Category: | MySQL Server: Documentation | Severity: | S2 (Serious) |
Version: | 5.1-bk | OS: | Any |
Assigned to: | Paul DuBois | CPU Architecture: | Any |
Tags: | regression |
[13 Dec 2007 12:55]
Domas Mituzas
[13 Dec 2007 13:21]
Heikki Tuuri
Domas, this is the correct behavior. I think Guilhem or someone added a warning print if someone uses READ COMMITTED in 5.1 with statement-based replication. Users should in general move to row-based replication in 5.1, because it is theoretically less buggy that statement-based. Regards, Heikki
[13 Dec 2007 15:26]
Domas Mituzas
it has to be documented as incompatible change, if we want to tag it as 'not a bug'.
[13 Dec 2007 16:01]
Heikki Tuuri
Paul, please check that there are prominent notices in the 5.1 manual: if you use READ COMMITTED (which is equivalent to innodb_locks_unsafe_for_binlog in 5.0), you MUST use row-based binlogging. Regards, Heikki
[20 Dec 2007 0:55]
Paul DuBois
Changing to Server: Documentation/Verified because this is a documentation issue. Assigning to myself.
[20 Dec 2007 1:01]
Paul DuBois
Thank you for your bug report. This issue has been addressed in the documentation. The updated documentation will appear on our website shortly, and will be included in the next release of the relevant products. The update will appear here: http://dev.mysql.com/doc/refman/5.1/en/innodb-transaction-isolation.html