Waiting for Postgres 16: Cumulative I/O statistics with pg_stat_io | Hacker News Reader

Waiting for Postgres 16: Cumulative I/O statistics with pg_stat_io

pganalyze.com

168 pointslfittl29 comments

Does anyone know if 16 will bring back JSON_TABLE from Postgres 15? They initially had it in the beta, then at the last moment, removed it.

Not yet clear. A lot of the infrastructure needed to support JSON_TABLE et al in a cleaner fashion, has now been added.

Good to see this part of the performance problem space being addressed.

I see lots of similarities between the (newer) Postgres wait analysis logging/views and Oracle AWR/ASH. I suspect it was heavily inspired.

Inspired, or solving similar problems in a similar tool leading to similar solutions?

Indeed, shouldn't patents be checked to ensure there isn't any risk here?

No. Oracle can go bankrupt for all I care. Criminal business practices.

Wait events are perhaps inspired by oracle's stuff, at least indirectly. It was developed by Robert Haas and Amit Kapila, working for EDB. Traditionally a lot of their customers have migrated from Oracle.

The redesigned way of storing stats in memory and pg_stat_io aren't inspired by Oracle. Melanie, who authored pg_stat_io, has never used oracle. I haven't used oracle in ~15 years, and the only thing I remember is fighting crashes in connect_by (I influenced the design of pg_stat_io, and authored large portions of the shared memory stats stuff).

Edit: noted Amit's co-authorship of wait events

Was thinking of upgrading to 15. So should I wait for 16?

While waiting for PG 16 you can upgrade to 15.

Agreed. I also see PG 16 is scheduled for release later in the fall. Also, you don't want to put the first 16.0 version on a production system, rather wait until 16.1, 16.2 etc.

That's always for the best. So now that 15 has had some time to settle, it's a good time to adopt.

Great opportunity to test your db backup process. :)

Because PostgreSQL major version upgrades are hard and for large databases, requires some level of downtime. Given that that major version releases are on an annual release cycle, staying evergreen is really really annoying.

Under some elementary scenarios (and using some Postgres offerings), you can do a full major upgrade using logical replication (even with replicas). It depends on your situation, generally I agree, any sophisticated use of Postgres (lot of views, materialized or not, fast changing schema, large objects, and so on) makes upgrades a pain and is generally somewhat of an unresolved problem in my opinion.

It would be nice if the Azure and AWS Postgres teams could get off their asses and release version 15 already. It has been out for several months now, and crickets.

Homebrew too, but at least they have the excuse that it is open source and unfunded.

Stripminers have to rest too

Same goes for GCP. Normally they release it before Christmas but version 15 still hasn’t been released.

Homebrew does have Postgres 15. https://formulae.brew.sh/formula/postgresql@15

It’s keg only, which shouldn’t matter, except that it does if you want to add any extensions. For example, postgis can’t be added to Postgres 15 using normal brew commands (you can if you edit it and build from scratch and link manually, but that’s kind of a nightmare). Everything is set up to treat PG14 as the default, because that is the one that is not a keg.

I realise it may not be possible for many people due to their organisational requirements, but I think self-hosting is a great option when possible, and the backup story is great with pgbackrest.

Easy to upgrade when you're ready.

Just use Crunchydata. Runs in any of the clouds and gives you better options.

It can be a good thing to wait a minor version or two before upgrading to the latest major version. There's much more confidence in the release that way, some companies even have a policy along those lines.

15.1 came out in November.

15.2 came out earlier this month

https://www.postgresql.org/about/news/postgresql-152-147-131...

Unrelated with the topic but pganalyze seems great (never heard about it before).

That's how you make a good home page.

OP here - thanks for noting!

We've recently invested effort to make the homepage better (added the animation, revised the messaging, etc), so that makes me happy to hear & will forward to the team as well.

Hey I added your blog to my RSS feeder and it seems that not all articles from your site are in the feed. I can’t inspect feed.xml from mobile to confirm if it’s a problem on your end or mine though

Good catch. Thats actually intentional at the moment - we don't put the weekly "5mins of Postgres" episodes in our RSS feed.

Primarily since we syndicate the feed to Planet Postgres, and video would go against the rules. I'll see if we can split up our feeds somehow.

Reply on news.ycombinator.com