Thank you for the explanation. Pavel On Tue, May 21, 2013 at 10:54 PM, Kristian Nielsen <knielsen@knielsen-hq.org> wrote:
Pavel Ivanov <pivanof@google.com> writes:
Could you please elaborate what kind of bug is that?
The slave connects to master with GTID. The binlog dump thread on the master starts scanning the last binlog file to find the start position. It finds the position for one replication domain, but aother domain is still scanning to find the right place to start.
Then the slave looses the connection to the master, due to some network problem. It automatically reconnects. Then the slave reconnect will connect with _both_ domains at the binlog offset reached before the connect. Thus, if there were still events in the second domain to be scanned and skipped at the point of reconnect, those events might be duplicated.
(This happens only with the automatic reconnect of the I/O thread, not with explicit slave thread restart).
- Kristian.