Re: replication_slots usability issue

On 2018-10-30 11:02:04 -0700, Joshua D. Drake wrote:
> On 10/30/18 10:52 AM, Andres Freund wrote:
> > On 2018-10-30 11:51:09 +0900, Michael Paquier wrote:
> > > On Mon, Oct 29, 2018 at 12:13:04PM -0700, Andres Freund wrote:
> > > > I don't think this quite is the problem. ISTM the issue is rather that
> > > > StartupReplicationSlots() *needs* to check whether wal_level > minimal,
> > > > and doesn't. So you can create a slot, shutdown, change wal_level,
> > > > startup. A slot exists but won't work correctly.
> > > It seems to me that what we are looking for is just to complain at
> > > startup if we find any slot data and if trying to start up with
> > > wal_level = minimal.
> > Right, we really should just call CheckSlotRequirements() before doing
> > so. I'll make it so, once I'm actually awake and had some coffee.
>
> Why not just disable the slot and report an INFO: line?

Because afterwards there'd be a slot with corrupted contents. Especially
bad for logical slots. Which might look ok, but crash in weird ways.

- Andres