this post was submitted on 05 Jul 2023
16 points (100.0% liked)
PostgreSQL
665 readers
1 users here now
The world's most advanced open source relational database
Project
- About (history)
- Docs
- Donate to PostgreSQL
- Wiki
- Planet PostgreSQL
- IRC
- Mailing lists:
- pgsql-announce
- pgsql-hackers (developers)
- pgsql-general
- pgsql-jobs
- User Groups
Events
- SEAPUG Summer BBQ, 6 July in Seattle
- SFBA PostgreSQL Meetup, 12 July
- Chicago PostgreSQL Meetup, 19 July
- PGDay UK 2023, 12 September in London
- PGConf 2023, 3-5 October in New York City
- PGDay Israel 2023, 19 October
- PGConf.EU 2023, 12-15 December in Prague
Podcasts
Related Fediverse communities
- c/SQL on programming.dev
- #sql on Mastodon
- #postgresql on Mastodon
founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
Considering clusters need a load balancer in front of them, I'm going to say only through add-ons.
You can shard a table on the same server (it's called partitioning).
Not something I have done, haven't needed to yet.
Beyond that, it's pgcat I believe, in which query sharding is still experimental.
HAProxy and PGBouncer can do load balancing for reads, but ultimately the cluster is single main for writes.
Supabases Supavisor is another beta release option for postgres connection routing, but it is VERY early in development. I think they are dog-fooding it tho. Certainly one I am keeping an eye on!
However, that is my limited experience opinion.
Happy to be told otherwise, hopefully I start a discussion!
thank you! I'm too tired to contribute to a discussion right now ๐
been experimenting with postgres to hopefully move to it eventually.
I use mariadb at the moment and they've been releasing pretty unreliable updates in recent years. Galera was nice with it, but had way too many issues.
If you're looking for a mysql/maria drop-in replacement, then you may also want to look at tidb
postgres can do it, but it's kind of presented as lego blocks for a HA cluster and often requires making your own tooling if you want sophisticated features like placement index instances and rebalancing
query optimization and the similar structure to SQL server with schemas pulls me towards postgres, got another system using SQL server with upwards of millions of records and need it to be fast and reliable