Well, first of all, it doesn't ensure that. There's no communication between memcached nodes in a cluster, which is what makes it scale linearly.
Second, you shouldn't worry about a node filling up. It's a cache, not a data-store, the least recently used items will be evicted when you fill up a node. Third, since all keys are hashed, you'll find that your items will be evenly distributed so unless your nodes have different amounts of memory, they will fill up at approximately the same pace. /Henrik On Sat, May 30, 2009 at 07:05, scranthdaddy <[email protected]> wrote: > > The documentation states "When doing a memcached lookup, first the > client hashes the key against the whole list of servers. Once it has > chosen a server, the client then sends its request, and the server > does an internal hash key lookup for the actual item data. " > > My question is how then does memcache ensure that a node in the > cluster doesnt fill up? >
