Re: logical changeset generation v6.8

On 2013-12-16 00:53:10 -0500, Robert Haas wrote:
> > Yes, I think we could mostly reuse it, we'd probably want to add a field
> > or two more (application_name, sync_prio?). I have been wondering
> > whether some of the code in replication/logical/logical.c shouldn't be
> > in replication/slot.c or similar. So far I've opted for leaving it in
> > its current place since it would have to change a bit for a more general
> > role.
>
> I strongly favor moving the slot-related code to someplace with "slot"
> in the name, and replication/slot.c seems about right. Even if we
> don't extend them to cover non-logical replication in this release,
> we'll probably do it eventually, and it'd be better if that didn't
> require moving large amounts of code between files.

Any opinion on the storage location of the slot files? It's currently
pg_llog/$slotname/state[.tmp]. It's a directory so we have a location
during logical decoding to spill data to...

Greetings,

Andres Freund

--
Andres Freund http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services