This is a cool idea, but I wonder how useful it is when running a large number of memcached servers. Presumably one would have to connect to each one and send the pget or pdelete commands?
I currently achieve the pdelete effect in code by using a predefined prefix and then "abandoning" the keys when they need to be deleted simply by having the prefix change. i.e. 123:getUser:10 becomes 124:getUser:10. Technically the data for 123:getUser:10 is still in memcache, but nothing will ever read it and it will eventually get flushed out. I don't know of any easy way to do something like pget in a client, and I can think of a couple instances where it would be useful, but it would also be incredibly easy to end up requesting way too much data, and again with having to connect to each memcached server, it might still be more hassle than it's worth. Jay Josh Dybnis wrote: > memcached-prefix is an experimental fork off of the memcached 1.3 > development branch. It adds commands pget and pdelete that operate on > ranges of keys having a common prefix. The new commands can be used as > a simple namespace mechanism. It also adds a memcachedb compatible > rget command. > > Performance is very close to the standard memcached (see the > benchmarks on the project page). Space usage is also roughly > unchanged. > > Project page: http://jdybnis.github.com/memcached/