Re: Replication - always broken when poweroff one of the nodes

[TOMASZ SERWANSKI]

..and what is interesting:

1) created database 'tmp20' on first node
2) confirmed that database is visible on both nodes
3) deleted database from second node
4) confirmed it disappeared from both nodes
5) crash/poweroff on one of nodes
6) after restart, slave's status complains that:

Last_SQL_Error: Error 'Can't drop database 'tmp20'; database doesn't exist' on query.

so it looks like in binlog there are kept logs from already committed transactions, looks to me like an inapropriate behavior; could anyone advice how to approach it?

regards
Tomasz