Comment by taneliv

While Jepsen (and this article) is focused on behavior under node failure and network partitions, this caught my eye:

> It also exhibits Stale Read, Lost Update, and other forms of G-single in healthy clusters

This looks like quite a fundamental issue.