How does memcached cache?

Question

I have four large objects that I'm caching using Memcached. I've allocated 200MB to Memcached.

In my Python code, I cache the items like so:

cache.set('object_1', object1)  # fourth biggest in size
cache.set('object_2', object2)  # second biggest in size
cache.set('object_3', object3)  # third biggest in size
cache.set('object_4', object4)  # biggest in size

In my cache, I have access to object_1 and object_3 but not object_2 and object_4 - Is Memcached making a call on its own as to which objects to cache and which to ignore based on the available memory?

Answer 1

There's a good explanation of this here: Memcache Internals

"As you might already know, memcache will delete old data to make room for your new data, but it needs to know which data-object can be deleted. Is it the largest one so we instantly have a lot of room? Or maybe it’s the first one entered, so you get a FIFO based caching system? Turns out, memcache uses a much more advanced technique called LRU: Least Recently Used. In a nutshell: it deletes the item that isn’t used for the longest period of time. This doesn’t have to be the largest object though. And it even doesn’t have to be the first object stored in the cache."

Hope that helps!

How does memcached cache?

Follow this question

Related questions