Failover to cached_db

[jazzband]

Failover to cached_db

[teferi]

That doesn't seem like a good idea to me. It looks like this kind of failover operations should be executed by django itself, not by a cache backend.
Also the whole idea seems strange. Why would you want to failover to another backend, rather than to a redis-slave (which can be configured to be read-write) for example?
Also have you considered using tools like haproxy for failover?

[mauricioabreu]

Imagine if your Redis (server itself) turns down. You should have 2 servers with 2 Redis running. Am I missing something?
I will consider haproxy. Thanks for your tip, @teferi .

[niwinz]

Hi!
I really agree with @teferi ! But I have done some experiments in this area on some experimental client: https://github.com/niwibe/django-redis/blob/master/redis_cache/client/experimental.py#L52

It can be adapted for use an other backend instead of using an other connection string.

In any case, consider @teferi recomendations.

[mauricioabreu]

Thanks for your feedback. I am closing this issue because I agree with you both.
And yes, I will consider.