Rod Taylor wrote:
>>> Exactly what "replication work" do you mean?  One table? All tables being copied?
>>> In my situation I have 6500*5 + 100 tables to copy.  No way is that going to be
>>> completed in 5 minutes no matter that the tables are small.  (And no
>>> I did not design the schema :)
>> That isn't the initial COPY, for sure, because none of ours will
>> finish the COPY in 5 mins, either.  I suspect Rod means that there is
>> a time limit on how long a snapshot-application process runs.  It
>> can't be 5 mins, though, or we'd run into this after the initial
>> COPY.  Chris, can you shed more light on this one?
> 
> Not only that, but please note in my followup message it often happens
> in the opposite direction of the data flow.
> 
> Node 4 subscribes to datasets on Node 1.
> 
> Node 1 will (once every two months) result in hundreds of connections to
> Node 4.

Do you have a watchdog running for the node?  That's the exact pattern
of behavior you see when a watchdog with a 5 minute sleep time starts
killing the slon.

This will happen if the last confirmed event on the target node
(confirmed from the master, I think) is older than that the default time
in the watchdog (older than 20 minutes by default).

-- 
Brad Nicholson  416-673-4106    bnichols at ca.afilias.info
Database Administrator, Afilias Canada Corp.