Hi,

I have simple setup with one master db and 2 slaves. I have a stored
procedure that gets executed before insert on one of the replicated
tables. Everything propagates automatically fine, except for one small
problem - the slave nodes have 2 rows entered for every 1 on the
master - probably because the trigger gets fired there too.
I've added the trigger and procedure manually (that particular table
is not used atm).
The procedure checks if a row is already in the table and if so -
updates it (instead of doing and insert), in addition to that it
inserts one row to another table (which is replicated).

The easiest way would be to disable the trigger on the slaves, the
problem is that if master dies my application tries to write to the
slaves (there is a system in place that promotes one of the slaves  to
a master role).

How do you usually solve this sort of problem?

kind regards
Pshem