Memcache auto discovery #548
Conversation
Nerdyvedi
force-pushed
the
elasticache_auto_discovery
branch
from
960faab to
ea2e6da
Compare
Nerdyvedi
changed the title
Nerdyvedi
changed the title
jogo
left a comment
jogo
left a comment
There was a problem hiding this comment.
Idea looks good a few inline comments plus, unit test coverage and updated docs (including the changelog) would be great.
Thank you for the contribution, and sorry for the slow review.
| return True | ||
| return results[0] == b"OK" | ||
|
|
||
| def auto_discover(self): |
There was a problem hiding this comment.
is this command useful in the base client, where I assume we only have a single memcache ip port tuple, or is this only useful in the hash client? If only Hashclient lets move it there
There was a problem hiding this comment.
In my experience, it makes a sense to utilize HashClient alongside it, as it allows for connecting to multiple nodes simultaneously or you can connect to cache cluster with Client by using configuration node, ex. mycluster.fnjyzo.cfg.use1.cache.amazonaws.com.
As I saw here's not much activity for half year, but I would be happy with support of this feature for one of mine project. Could I handle it if nobody mind?
pymemcache/client/hash.py
Outdated
| Args: | ||
| servers: list() of tuple(hostname, port) or string containing a UNIX | ||
| socket path. | ||
| If enable_auto_discovery is set, just a configuration endpoint would suffice |
There was a problem hiding this comment.
It would be great to point out this is an AWS elasticache specific feature
Nerdyvedi
force-pushed
the
elasticache_auto_discovery
branch
from
ea2e6da to
0fe0600
Compare
3 participants
In Hash client, if instead of a list of nodes, just cluster configuration endpoint is passed , This adds an option to fetch nodes from the endpoint
Resolves #534