Hi.  Ran into this scenario:  replication lag increases from seconds
to minutes (or tens of minutes) while doing a pg_dump of the entire
database on the slave.  Per Steve Singer's suggestion I excluded the
slony schemas from the pg_dump and replication lag stayed within
normal paramaters.

May I offer an addition to the "Best Practices" page for both 1.2 and
2.0 documentation sets:

If you pg_dump your database from your Slony slave, avoid dumping your
Slony schemas or else pg_dump's locking will compete with Slony's own
locking which will slow down Slony replication for the duration of the
pg_dump.  Exclude the Slony schemas from pg_dump either using "-n
public" to dump the default schema only (assuming all your data is
there), or, if you use multiple schemas, use pg_dump
--exclude-schema=schemaname to specifically exclude your Slony schemas
by name.

Best,
Aleksey