RFD 1067: CockroachDB, reranked against FoundationDB
Problem
RFD 1020 picked CockroachDB before sibling weftspun repositories built and benchmarked FoundationDB, its Relational Layer, and mvsqlite. One of them, h2o-bench-tpcc, picked FoundationDB over CockroachDB, and called the CockroachDB fork this project pins “a dead engine.” Does that verdict carry over here.
Decision, as published and now retracted
This RFD published “keep CockroachDB, RFD 1020 stands.” That decision no longer holds. weftspun/cockroach-local is archived and out of default.xml, and datasource-foundationdb and datasource-store are forked in on side 6.
The retraction is narrow. Two of its three reasons were answered by events rather than found wrong, and the third never argued for moving. The dead-fork risk this RFD called open work is the part that aged badly. Today arrived.
Abandoned here means the decision was in force and has been reversed. It does not mean nobody acted on it.
See DETAILS.md for the three reasons in full, what answered each, the two questions this leaves open, and the evidence from each sibling repository.