Pat Maddox wrote:
> Sounds like theoretically it could matter, but in practice it doesn't.
> I'd like a more definite answer though.
Well, most of the locking issues with Slony seem to be with
administrative commands (setting up a replication set, altering it)
which require taking locks. If your application(s) lock tables in the
order C,B,A and slony in A,B,C then they can deadlock waiting on each
other. This is a problem you'll face any time you have two sets of
exclusive locks interacting.
I'm fortunate in that the systems I deal with all allow for some
downtime in application access, so I just schedule slony changes for
these periods.
--
Richard Huxton
Archonet Ltd