I suggest writing a very simple shell script to run "SHOW SLAVE
STATUS" and output it to a file every 15 seconds, to verify your
script is working. I have never seen MySQL give a bad lag time for
replication -- it's always been accurate for lag time, or 0, or NULL.
As a bonus you could then use the data from the log files and
positions to actually calculate how far behind it was, and file a bug
report if indeed the server is wonky.
On 5/19/06, Martijn van den Burg <m_a_r_t_ij_n@stripped> wrote:
> > That is weird. If it only lasts a couple of seconds, how are you
> > monitoring it to find out what the lag time is?
> I've written a replication monitor script using Perl::POE, which checks
> replication lag every 15 seconds or so (can't check the exact interval now -
> weekend has begun here).