Let's say I'm setting up a new slave. I start with a mysqldump after running mysql -e "STOP SLAVE", transfer that and restore it to the new server.
If I didn't capture the BINLOG filename and position after the STOP SLAVE and before running the mysqldump then what happens if I simply use a previous/older binlog filename (verified to be present on the master and older than the backup), with an position of zero?
It seems to just force the new master to replay all of the events to the new slave, obviously including a number of INSERTs, UPDATEs, and DELETEs which were already represented in the backup (those which happened during the overlapping time period).
Is there anything wrong or risky about that? Is there a downside other than the extra wasted time and CPU cycles (replaying perhaps a day's worth of overlapping transactions)?