### Check for existing issues

- [x] I have searched the existing issues and checked that my issue is not a duplicate.

### What happened?

Setting `REDIS_CLUSTER_NODES` causes the application shutdown to fail (see logs).

Observed on litellm:v1.83.3-stable. However, I think the codepath is still active on `litellm_internal_staging`. I'm a bit new here, so forgive me if I haven't followed the path correctly.

Here's my theory:

In _redis.py: `if "startup_nodes" in redis_kwargs: return None` means `get_redis_connection_pool()` returns `None` if `REDIS_CLUSTER_NODES` is set. (https://github.com/BerriAI/litellm/blob/8bca05d311ab3ce9b0d213a3bcd2d9d503fc7477/litellm/_redis.py#L668-L669)

This means that `RedisCache.__init__` sets `self.async_redis_conn_pool = None` (https://github.com/BerriAI/litellm/blob/8bca05d311ab3ce9b0d213a3bcd2d9d503fc7477/litellm/caching/redis_cache.py#L239)

However, on shutdown, `disconnect` is called, which does not expect `async_redis_conn_pool` to be `None`: https://github.com/BerriAI/litellm/blob/8bca05d311ab3ce9b0d213a3bcd2d9d503fc7477/litellm/caching/redis_cache.py#L1278

An `if None:` guard in `disconnect` should fix this. I can provide a PR if you agree…