On Mon, May 14, 2007 at 02:33:45AM +0400, Dmitry Koterov wrote:
> They are analogs of LISTEN/NOTIFY, but use standard IPC instead of
> pg_listener. Then I:

I'd be astonished if pg_listener were your main problem.

> So, if anybody reads from sl_log_status, it blocks until anything is written
> to sl_log_status. Seems it will be fully transparent for Slony, but
> decreases the amount of polling queries. (It is an idea only, practical
> implementation may need to be more detailed.) What do you think about it?

It sounds like a way to make everything a lot slower.  

One way you could improve your notification about whether a slave is
up to date enough for you is to add a sequence.  When you start your
transaction, you select nextval() from the sequence.  You can poll
from time to time on the replicas to see whether the sequence is >=
the value you have.  If so, that replica is "up to date enough" for
you.  But my real feeling is that Slony is the wrong tool for the job
you're trying to use it for. 

A

-- 
Andrew Sullivan  | ajs at crankycanuck.ca
When my information changes, I alter my conclusions.  What do you do sir?
		--attr. John Maynard Keynes