Re: Better support of exported snapshots with pg_dump

On 2014-09-01 21:54:24 +0900, Michael Paquier wrote:
> On Mon, Sep 1, 2014 at 6:30 PM, Andres Freund <andres(at)2ndquadrant(dot)com> wrote:
> > I was never convinced of the reasoning in that thread. Possibly things
> > have changed enough now that logical decoding is in core...
>
> Well, the test case I got in mind is only for taking a dump using the
> latest state of a replication slot and not the snapshot export itself.

I don't think what you're proposing is really possible. Could you
describe it in a bit more detail?

> So what about the following: we let the user specify a slot name with
> pg_dump, and take a dump using the latest snapshot that this
> replication slot has reported to a user.

There exists no snapshot sufficient for user data after slot creation.

> > I don't think they change anything here. The problem is the, pretty
> > fundamental, problem that you need to know a relation exists before
> > executing a LOCK ...; on it. During that time somebody can change the
> > schema.
>
> Doesn't this window exist as well with parallel pg_dump?

Yes. I didn't say those reasons were convincing. The window is quite a
bit smaller though. With the exported snapshot from CREATE REPLICATION
SLOT it could convinceably be hours.

Greetings,

Andres Freund

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