> Running 1.0.2, I have a replication cluster with one set defined, with a
> slon running. I don't have any subscribers currently. I would expect the
> local slon to just clean out sl_log_1, but that table keeps growing. I
> don't see any errors in the slon log. Is my understanding correct that
> the log table should be getting cleaned out, or do there have to be
> subscriptions for some reason for this to happen?

Version 1.0.2 has an issue that it doesn't do enough cleanup of the
sl_confirm table.

I'll bet you used to have some nodes that are gone now?  Those nodes
presumably left entrails in sl_confirm that are causing the cleanup code
to think it should hold onto the sl_log_1 data until the (now gone) node
comes to pick up the data.

This is addressed in the FAQ under the following question...
http://linuxfinances.info/info/faq.html
Q: After dropping a node, sl_log_1 isn't getting purged out anymore.

The resolution involves deleting the obsolete sl_confirm entries.