- The query language is JS, and allows both 'shape of data' and functional queries
- There's live change feeds (which means the DB, being the source of truth, takes the role of initiating change messages)
- RethinkDB has an excellent reputation for being able to get the data back after you save it.
Basically it's like Mongo but not (insert adjective).
We've been using it in production for 2 years at CertSimple and have been very happy. Previous experience is Mongo, GAE data store, and various ORMs pointed at SQL. The docs are great, the defaults are safe, and doing new things is easy.
> RethinkDB has an excellent reputation for being able to get the data back
Uh. It's not always true, although not in the sense you've meant it (no complaints about storage reliability).
I don't know if my setup is broken or I did something stupid, but for me rethinkdb-dump saturates the CPU to the max and the only thing that keeps the machine from choking to death with LA1 going over 100 is resource limit on the database container. Trying to back things up results in random connection drops and timeouts. I gave up on trying to back up the database online.
And that's a very small database (75GB on disk, 12GB as uncompressed JSON, 2.5GB as a tarball), on a reasonably powerful machine. It's a single node, though - I thought I'd "upgrade" to a cluster at some point but it's way too early.
- The data can be structured
- It can also have relations
- The query language is JS, and allows both 'shape of data' and functional queries
- There's live change feeds (which means the DB, being the source of truth, takes the role of initiating change messages)
- RethinkDB has an excellent reputation for being able to get the data back after you save it.
Basically it's like Mongo but not (insert adjective).
We've been using it in production for 2 years at CertSimple and have been very happy. Previous experience is Mongo, GAE data store, and various ORMs pointed at SQL. The docs are great, the defaults are safe, and doing new things is easy.