4.2-roadmap.md

Redis 4.2 roadmap

1. Redis Cluster

• Speed up key -> hashslot association. Now makes RBB loading 4x slower when there are many small keys.
• Better multi data center story
• redis-trib C coded and moved into redis-cli
• Backup / Restore of Cluster
• Non blocking MIGRATE (also consider not using 2X memory)
• Faster resharding
• Bug fixing and stress testing to bring it to next level of maturity
• Potentially a cache-only mode, slaveless, masters automatically reassign slots after failures

2. Streams data structure
3. Redis Modules improvements:

• Thread-safe contexts
• Thread-safe key locking
• Cluster API and reimplementation of Disque as a Redis module

4. Listpacks replacing ziplists

super news. what is the tentative date ? 4.2 is in beta.
Interested to see this feature - Non blocking MIGRATE (also consider not using 2X memory).
Particularly to do away the need of 2X memory

Hi, I've designed & implemented an asynchronous data migration on Codis (based on Redis 3.2.8), nearly 7x times faster than the synced implementation.

Here's the post on reddit.

cool to see Disque on the list 👍

disque is cool !

Cannot come soon enough!

Finally redis-trib directly into redis-cli, better cluster automation & disque "official" release!!!

I cannot wait to use the "slaveless" mode. That will make redis operations a lot easier.
Truly exciting!

Great to see that goal. Looking forward to see Redis as a message broker !!!

Hi Salvatore,

As you might know, we are heavy users of Redis at Pandora and have our own tooling / sharding / etc... around it. We're very interested in 4.2 especially for the multi-dc story and will possibly migrate everything to Redis Cluster on that version. I have a couple questions to help us in our planning for this year:

1. Is there a tentative date or maybe a very rough one for a Redis 4.2 release?
2. Do you have any more details at this point about what the Multi-DataCenter support will be? Would we be able to easily run a read-only copy of the entire Redis Cluster in another data center (with a unidirectional replication), or is it going to be a live-live (master-master) setup?

Thanks!

thanks