> Yesterday I set up 4 new MySQL 4.0.22 replication slaves on 2 FreeBSD
> 4.10-R machines (via mysqld_multi). We have 10 other identical slaves and
> I followed the same procedure as always when seeding them. The new slaves
> were working fine, however when I issued SHOW SLAVE STATUS the
> Relay_Master_Log_File field was empty. I reseeded one of them [with
> freshly created seeds] a couple of times with the same result every time.
> Then this morning I woke up and found that Relay_Master_Log_File was
> I haven't been able to find anything about this online yet, and I am
> wondering if anyone else has experienced this? The reason why I caught
> this is that our monitoring software relies on that field being populated
> to calculate replication delay.
AFAIK, the log file name is not known by the slave unless it either
receives a rotate log event (go to next log) or you have started it
against a particular log file. If you start replication "from the
beginning" (not specifying a log file) against a master, the first
filename is not passed down.
In order to force a filename to be displayed you could either: FLUSH
LOGS on the master, or specify the log file name in the CHANGE MASTER on
Technical Yahoo - MySQL (Database) Geek
Desk: 408 349 5104