Docs on distributed memcached setup are not clear

[ciaran....@activeenergy.ie]

https://docs.djangoproject.com/en/2.1/topics/cache/#memcached

Particularly the multiple servers section, which does not detail if all memcached daemons should be configured to listen on all interfaces i.e. 0.0.0.0

So just as a sanity check:

Scenario 1: (memcached installed naively ( i.e. listening on localhost:11211) on multiple ubuntu servers)

* Dev sets up her CACHE_LOCATIONS = ['server1:11211', 'server2:11211'] but is none the wiser that there is no interoperability between servers, meaning roughly 50% cache misses in totality.

* Each django process cycles through each location, only localhost is successful, no warning logs generated

Scenario 1: (memcached installed and configured to listen on 0.0.0.0:11211) on multiple ubuntu servers)

* Dev sets up her CACHE_LOCATIONS = ['server1:11211', 'server2:11211']  and django add/sets keys on all memcache servers successfully

* Each django process cycles through each location, adding/setting keys as it goes

So, am I right in thinking that the docs assume Scenario 2, but Scenario 1 is a plausible setup, it would just mean lots of misses, but then a pro of no network overhead.

Am I missing something?