Want to build a RasPi cluster shaped like an old-school Cray supercomputer? 'course you do! You can find the greyprints and see how to assemble it here.

Notes and scripts for setting up (yet another) Raspberry Pi computing cluster. One master, at least one slave to do the actual work. The master implements a certain amount of infrastructure for the rest of the network. Includes greyprints for 3D printing a rack for the units. Uses k3s and Docker.

Redis is another key-value datastore (a distributed hash table, if you like) that can be used for either caching of often used objects or a database to build applications on top of. Can handle just about any serializable data structure, from strings to sets. Clusterable - uses master/slave replication. Written in C but has bindings for any programming language you're likely to know.

A distributed relational database that uses SQLite as its back end. Eventual consensus of all nodes that relies upon quorum. Automatically elects leaders in the network of servers. More lightweight than MySQL or Postgres. Has a REST API built in. Fully transactional. Hot backups possible. Uses SQLite in "all in memory" mode by default for efficiency, but can maintain on-disk database files with a command line option (the on-disk transaction log can be used to repopulate the in-memory database on startup). Written in Go.