database_guy

At work I have a multi-petabyte cluster in MongoDB Atlas with no issues. The managed service is fantastic; I'm going to need some citations on how it degrades after a few terabytes. I don't really like MongoDB's document-model-everywhere approach but I have to respect how well their managed service works.

It started as a proprietary database. Apple acquired it and then open sourced it eventually.

I lamented this exact situation in a post I made last week. I'm with you on this completely.

Fair point re: "someone has to start the project". I think Yugabyte is neat, and I want to invest in it as a platform, but I can't help but look at the project and feel like it's not set up for success should the Yugabyte the company fail. I hope I'm wrong and in 10 years it looks something like the Postgres community.

I also did eventually find the docs for building the database. It just feels odd to have a roadmap in the readme but no actual information for getting set up on development. It's one of those things (and I grant it's a very small thing) that makes me feel like the platform isn't really for the community.

I really liked RethinkDB (maybe even loved?), but its demise is one of those cautionary tales that makes me wary of these single-corporate-sponsor databases.

Cassandra does feel the best here to me. Thanks for the experience report, maybe I should test out how it runs when confined to somewhat small scale.

My point is that for small scale for my uses the differences between these databases are mostly academic. I am fully confident in my ability to model data around their various constraints and get them to do pretty much anything. I also understand why their HA functions differently. To be clear, there are projects that, governance/development community aside, hit all the right spots technology-wise for what what I'm looking for. These projects just seem to have far less healthy communities to me than others and I'm curious why that seems to be the case.

Perhaps I'll have to test it out at larger data sizes. The performance document makes it sound like this is here-there-be-dragons territory. I'd love to see how the database holds up at 100GB. Rqlite is a great project and I'm glad it exists!

I agree, if you want to offload a lot of this to Kubernetes there seems to be quite a bit more out there. Percona has some neat stuff.

Yeah, sounds like most folks agree this doesn't exist simply because it's very challenging to implement.

Rqlite is an awesome project but it's not intended for use above a couple GB, like etcd. Sorry, dataset size is what I meant by cluster-size which was a confusing way to put it. You're right, I don't care about the number of nodes in the cluster.

I have contributed code to projects "owned" by one of these types of organizations and after 8 months of waiting for my (very small) patches to be reviewed because I wasn't internal, it just didn't feel like healthy Open Source. Calling them Fake Community Projects as a term works just as well for me.

OpenSearch could be a contender (though I did not have fun running Elasticsearch back in the day). I will keep watching it and see how development goes, it's a good call-out.

I am very interested in any project that implements multi-active SQLite clusters with no cluster size limit, do you have any links for me?

From my research I think Cockroach, Yugabyte, and Scylla are all interesting and solid products depending on the exact API/guarantees you're looking for. For products I don't even have the option to self-host I think DynamoDB and Cloud Spanner are great products, probably among others.