State of redis_failover

[yolk]

Hi,

is redis_failover still maintained? Is anybody using this still in production? I ask because the last coments on the issues #62, #64 are one year old now, but the fork from @fanhattan seems to see a lot of activity (and etcd-support which would be awesome). Maybe someone (@arohter, @ryanlecompte, @mperham) can comment on any further plans?

Thanks!

[ryanlecompte]

@yolk, this gem is no longer maintained by me directly. I still keep an eye on incoming issues and pull requests, however I'm no longer actively adding new features and fixes myself. I no longer use Redis on a daily basis and don't have the bandwidth. I'd love to see a PR from @fanhattan with all of the fixes and features that were added in his fork. Also, at the time Redis Cluster wasn't really around. I haven't checked, but I'd imagine it has matured since the time this gem was originally created. It may be a better solution at this point than redis_failover.

[ThomasAlxDmy]

@yolk @ryanlecompte At Fanhattan we are heavily using redis_failover in production. We believe that it is still more mature and more flexible than the redis-sentinel model. I recently added the etcd backend support because we were not fully satisfied with zookeeper. And then refactored the code quite a bit. I also fixed few legacy issues we found during our extensive failover tesing. Issues #62, #64 are fixed for instance.

We are still kicking the tires, but should feel comfortable merging soon; but we ARE currently using our etcd code in production on multiple clusters.

[ryanlecompte]

@ThomasAlxDmy, that's great to hear! Definitely please do submit a PR with your fixes/improvements once you're read. Your comment about the current status of redis-sentinel model is also interesting. Thanks!