3 ms·
I'd still rather use PostgreSQL for my data needs. psql is quite nice too as an interface, and "setup and forget" replication for everything in the cluster and
by megous 6y ago
I'd still rather use PostgreSQL for my data needs. psql is quite nice too as an interface, and "setup and forget" replication for everything in the cluster and remote access makes up for slightly harder deployment and upgrades.
Knowing that I can do anything and the change hits 3 hard drives in a span of few seconds on various machnies is nice. Tolerance for arbitrary down time of master/slave clusters is also nice.
But yeah, for quick one-off data munging, sqlite looks quite nice.
- benbjohnson 6y agoI agree that PostgreSQL has some great tooling around it. For SQLite replication, you might want to checkout Litestream: https://litestream.io/ https://litestream.io/
- megous 6y agoI know about litestream, it's not the kind of "setup and forget" I have in mind. But it's nice that it exists.
- benbjohnson 6y agoThat's good to know. I'm always trying to improve the developer ergonomics. What would you want to see added to make it "setup and forget"?
- megous 6y agoI don't think there's much to be done. I just like the fact that I don't have to think about replicating individual databases, regardless if I rename, delete, re-create them. It's being done at cluster level, it's one time setup and I can forget about it.