> Which means that the data that hits a single SQLite instance now needs to be replicated to all the others. My initial reaction: Yeah because ... SQLite is n

> Which means that the data that hits a single SQLite instance now needs to be replicated to all the others. My initial reaction: Yeah because ... SQLite is not a database server - it's an awesome API on top of a flat file? Do you want a replicated database? Then use a replicated database. Then looking at the list of solutions (rqlite, BedrockdB, dqlite, ChiselStore, LiteFS) ... I guess people really do want to make it replicated. Why not use a SQL server? Also: I was using QEMU on an old 32-bit Dell server that was before Intel's VM efforts. It worked decently. :)