Promoted master will still show previous slave status information, causing the incorrect binlog and position to be written to --mysql-master-status-file

[Todd Roman]

Please note - I was given some incorrect information. Percona does not have a behavior that is specific to this case. What can occur is that someone does a "reset master" on a node that was previously a slave, without doing "reset slave all". This will result in 'show slave status' returning data.

One option is to force master behavior, using a --force-master command line option and not try to get slave status information first. Another is to 'show slave status' and check that 'Slave_IO_State' is null.

I'm doing the latter for now, but, procedurally, will change things so a 'reset slave all' is done when the slave is promoted to master in a 1-way replication scheme.

306c306
< $mysql_logfile = $slave_status->{Slave_IO_State} ? $slave_status->{Master_Log_File} : undef;
---
> $mysql_logfile = $slave_status->{Master_Log_File};