> I'm in process of upgrading a master server from 4.0.24-log to 5.0.22-log
> in a single master-slave environment. I've previously upgraded the slave
> to 5.0.22-log and restarted replication without issue.
> The current master is running on RH9 and the slave is running on CentOS 5,
> which is what I'm trying to upgrade the master to, also.
> After synchronizing the databases, I've run 'reset master' and 'reset
> slave' on their respective servers, then 'change master to...' on the
> slave, but 'show slave status' always displays: 'Slave_IO_Running: No'
> And here is the log entry from the slave:
> 071113 9:16:19 [ERROR] Slave I/O thread: error connecting to master
> 'username@stripped:3306': Error: 'Lost connection to MySQL server
> during query' errno: 2013 retry-time: 60 retries: 86400
> 071113 9:17:35 [Note] Slave I/O thread killed while connecting to master
> The replication user has 'replication slave' privileges, and basically
> I've duplicated my.cnf from the old to the new master.
> I've done some Googling and searching of archives but not much luck.
> Fortunately, I've been trying out the master upgrade on a temp machine,
> and it's fairly easy to revert back to the current master server.
> But if anyone has thoughts or suggestions about what to try next, I'd be
> most appreciative.
Check the server_id on each server. Sounds like one or more of the
servers either doesn't have it explicitly set in my.cnf, or there is a
duplicated server ID somewhere.
Note that the default value doesn't count as "being set" in MySQL 5. It
has to be in the my.cnf file. That's my experience anyway.