Skip to content

Comment on A proposal for more reliable locks using Redisparent

Comments

It is not new at all, that's exactly why I wrote the blog post, since it is an actual use case. But it is a good idea to formalize a canonical model to do it with Redis in a safe way, and to show what are the failure modes when we resort to simpler designs (like failover to slaves).

As a related aside, I'd be mildly interested in Redis having an ability to produce and produce and set UUIDs. That sort of thing might also have some utility for locking structures/synchronization primitives, though it's not the original reason I'd thought having something like that might be nice.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.