Re: A Modest Upgrade Proposal

On Thu, Jul 14, 2016 at 2:29 AM, Craig Ringer <craig@2ndquadrant.com> wrote:
> Yes, I'd like that too. I'd also like to have fully parallized writeable
> queries right now. But we can't build everything all at once.

I agree.

> Before doing parallelized writes, things like dsm, dsm queues, group
> locking, worker management, and read parallelism were all necessary.

Yep.

> It's the same with cluster-wide management, dump and restore of replication
> state to re-create a replication setup elsewhere, etc. We have to build the
> groundwork first. Trying to pour the top storey concrete when the bottom
> storey isn't even there yet isn't going to work out. You've argued
> effectively the same thing elsewhere, saying that the pglogical submission
> tried to do too much and should be further cut down.

Absolutely.  I didn't mean to imply that the scope of that submission
should be expanded.  What I'm a bit concerned about is that maybe we
haven't given enough thought to how some of this stuff is going to
work.  Petr replied earlier with an assertion that all of the things
that I mentioned could be done using pglogical, but I'm not convinced.
I don't see how you can use pglogical to build a self-healing
replicated cluster, which is ultimately what people want.  Maybe
that's just because I'm not personally deeply enmeshed in that
project, but I do try to read all of the relevant threads on
pgsql-hackers and keep tabs on what is happening.

Suppose server A is publishing to server B.  Well, clearly, A needs to
have a slot for server B, but does that slot correspond precisely to
the publication, or is that represented in some other way?  How is the
subscription represented on server B?  What happens if either A or B
undergoes a dump-and-restore cycle?  It's just fuzzy to me how this
stuff is supposed to work in detail.

>> I don't understand this.  We add new DDL in new releases, and we avoid
>> changing the meaning existing of DDL.  Using function interfaces won't
>> make it possible to change the meaning of existing syntax, and it
>> won't make it any more possible to add new syntax.  It will just make
>> replication commands be spelled differently from everything else.
>
> Say you want to upgrade from 9.4 to 10.0 using the new logical replication
> features. How would that be possible if you can't add the required
> interfaces for setting up the downstream side to 9.4 as an extension?
>
> I think what we're leaning toward here is "don't do that". Tools like
> pglogical will have to carry that load until the Pg versions with built-in
> replication become the "old" versions to be upgraded _from_.

That may be true, but it's hard to tell whether that's going to be
feasible anyway without a fleshed-out proposal for how this is all
going to work.  If this can be made to work for upgrades from 9.4 with
only an extension, that would IMHO be worth trying to do.  But, just
for example, adding a replication set capability to 10 isn't going to
affect that one way or the other.  For upgrades, you'll want to
replicate the whole database.

> Ideally the new infrastructure won't have to make normal (non-walsender)
> libpq connections and will work entirely over the walsender protocol. That's
> not extensible at all, so the point becomes kind of moot, it just can't be
> used for downversion upgrades. Pity, but cleaner in the long run.

Yeah.  I'm entirely willing to leave downgrades to earlier versions to
extensions.  "Cleaner in the long run" has got to be a high priority
for core features; if we had not followed that policy in the past,
we'd have an unmaintainable mess now.

> It does make me wonder if we should look at extension points for the
> walsender protocol though, now we're like to have a future desire for newer
> versions to connect to older versions - it'd be great if we could do
> something like pg_upgrade_support to allow an enhanced logical migration
> from 10.0 to 11.0 by installing some extension in 10.0 first.

Maybe, but let's get something that can work from >=10.0 to >=10.0 first.

>> I'm concerned about dump-and-restore
>> preserving as much state as is usefully possible, because I think
>> that's critical for the user experience
>
> Right. See the pain points caused by our current dump issues like the
> brokenness around dumping security labels, grants, etc on the database its
> self. It certainly matters.
>
> The keyword there is "usefully" though. Replication sets: definitely useful.
> Knowledge about what peers we were connected to and what we were up to on
> those peers: possibly useful, if we have some way to meaningfully encode
> that knowledge, but far from crucial, especially since we can't actually
> resume replay from them without replication slots and replication
> identifiers we can't dump.
>
> It seems we were mostly crossing wires about different assumptions about
> what dump and restore would include.

Yes, that may be the case.

-- 
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company