mb4nck

spock and pgedge are cool, but they are NOT open source (and they try hard to make it look like they are).

It is under the Confluent license (with s/confluent/pgedge/), basically a source-available-but-Amazon-cannot-use-it kind of license.

OTOH the PolarDB specific changes seem to be contained enough that if you decide to run it in production, you can probably just apply most of the changes from the v11 branch yourself.

But I agree it's not a very good look to code-drop something on a .2 release when there's been 2,5 years of fixes.

I don't know, maybe you should.

CockroachDB seems to be a distributed database system written in Go which has implemented a Postgres query/wire protocol compatibility layer.

PolarDB is a Postgres fork actually using the Postgres codebase and extending it to a distributed database system. Maybe one day they can unfork because it's possible to implement PolarDB on top of Postgres as an extension and/or they contribute/get all their changes into Postgres core.

Looks like it, yeah.

Also, Postgres 12 introduced pluggable storage, which might help to implement a shared-nothing architecture without huge changes to vanilla Postgres (I haven't looked at how large their delta is)

But that's also true for e.g. the US and russian presidents, no? Except those actually hold real power.

In practise, most german presidents will just be way too old four years after they leave office to be a viable candidate, and it looks a bit absurd for them to be pushed back into the office, exactly because there's very little real power.

For the record, on Debian unstable:

  root@pg1:~# LANG=C apt install postgresql-12 
  Reading package lists... Done
  [...]
  0 upgraded, 32 newly installed, 0 to remove and 0 not upgraded.
  Need to get 59.7 MB of archives.
  After this operation, 240 MB of additional disk space will be used.
  Do you want to continue? [Y/n] ^C
  root@pg1:~# LANG=C apt install postgresql-12 postgresql-12-postgis-3
  Reading package lists... Done
  [...]
  0 upgraded, 105 newly installed, 0 to remove and 0 not upgraded.
  Need to get 104 MB of archives.
  After this operation, 418 MB of additional disk space will be used.
  Do you want to continue? [Y/n]

I think the addition of LLVM for JIT execution of plans added a lot to the Postgres install baseline (at least in distribution packages, you can compile a much leaner postgres yourself).

Notably, that feature got written by somebody from Salesforce (I assume on their work time, but not sure), and finished up by a Postgres major contributor. It might be the first major contribution from Salesforce outside of all the stuff Tom Lane did when he was working there.

There are no good multi-master solution with a HA focus available for Postgres, at least no open source ones.

The patroni project checks most of your other requirements: docker/containerized design with leader election/continuing operation via etcd or another external DCS. Not sure how easy upgrades are but I think the procedure is at least documented.

To the contrary, I think journalists often only get to advise on the headline, and the editors then choose/override one to their liking.

patroni builds upon etcd and (optionally) haproxy, two rather mature pieces of infrastructure (which can both be made HA on their own if SPOFs are to be avoided).

I understand where you are coming from, but having this kind of multi-server-who-is-master knowledge baked into Postgres itself will surely take another couple of releases, if it will be included at all.

Probably BDR (bi-directional replication, master-master like logical replication) will be there first, but whether the question is whether it will help a lot for local scale-out workloads (as opposed to glueing two datacenters together and allow transactions on both sides).

Note that starting from Postgres 10 (which this thread is about), you don't need to adjust wal_level and max_wal_senders (or max_replication_slots, for that matter) anymore. You still have to enable hot_standby=on on the standbys, though.

(and it is in general a good idea to keep the configuration the same as much as possible between primary and standbys).