Commit graph

63 commits

Author SHA1 Message Date
mats@romeo.(none)
62f8230938 BUG#26634 (Valgrind failure in tree: memory loss for memory allocated in rpl_utility.h):
Adding code to release allocated memory when tables_to_lock list is
cleared.
2007-02-26 17:44:55 +01:00
mats@romeo.(none)
3920f529ea BUG#23171 (Illegal slave restart position):
Third patch of the bug fix where the code for skipping events and for
executing events is factored out into three functions:
- shall_skip() to decide if the event shall be skipped and the
  reason for it;
- do_apply_event(), where the event is applied to the database; and
- do_update_pos(), which updates the actual relay log position and
  group positions.
2007-01-17 15:06:37 +01:00
kent@mysql.com/kent-amd64.(none)
1e3237fefa Many files:
Changed header to GPL version 2 only
2006-12-27 02:23:51 +01:00
mats@romeo.(none)
1fb9105ebd BUG#23171 (Illegal slave restart group position):
Second patch to fix skipping code. Moving relay and binary log 
position changing code from do_apply_event [old exec_event()] into
do_update_pos() and doing other changes necessary to support that.

Fixing a bug that can cause deadlock if rotating binary log when committing
a changes to a transactional table that is not inside a transaction and
cause a rotate log.
2006-11-10 15:10:41 +01:00
mats@romeo.(none)
a726fec6e7 Merge romeo.(none):/home/bkroot/mysql-5.1-new-rpl
into  romeo.(none):/home/bk/b18581-mysql-5.1-new-rpl
2006-11-03 21:11:58 +01:00
mats@romeo.(none)
83f8e17258 BUG#18581 (Creation of system tables recorded in binlog causing slave failure):
Not replicating the mysql database *at all* any more. All changes to
mysql tables are replicated by replicating the statements that do
the changes.
2006-11-01 09:49:03 +01:00
lars@mysql.com/black.(none)
c06f9991b7 Refactoring: Moved master info functionality into rpl_mi.cc to
make code easier to maintain.
2006-10-31 16:51:51 +01:00
lars@mysql.com/black.(none)
1e038dc882 Refactoring: Moved rli code to new file rpl_rli.cc. The idea being that rli
should be a separate module (i.e. a class) to make it easier to maintain the 
code, e.g. by having checks within the rli checking sanity of data and making 
member variables private.  This will also ease implementation of multi-source 
and, at least in my fantasies :), make it possible in some future to have 
separate replication servers.
2006-10-31 12:23:14 +01:00
petr@mysql.com
9669d3ae58 Merge pchardin@bk-internal.mysql.com:/home/bk/mysql-5.1-runtime
into  mysql.com:/home/cps/mysql/devel/5.1-logs-final
2006-06-07 23:58:08 +04:00
mats@mysql.com
4d0b8ce66f Plugging memory leak in row-based replication triggered by
test rpl_err_ignoredtables.
2006-05-09 12:30:06 +02:00
petr@mysql.com
fec70c97c9 WL#3153 "Split logs". Recommit with post-review fixes 2006-05-05 10:45:58 +04:00
mats@mysql.com
2884408c75 WL#3023 (Use locks in a statement-like manner):
Table maps are now written on aquiring locks to tables and released
  at the end of each logical statement.
2006-02-16 08:30:53 +01:00
lars@mysql.com
ad126d90e0 WL#1012: All changes as one single changeset.
This includes both code and test cases.
2005-12-22 06:39:02 +01:00