On Tue, Feb 16, 2010 at 10:22, Josh Carter <[email protected]> wrote: > On the memory side, I tried redis [1] at the suggestion of another list > member (thanks Gabe!) and so far it's great. It replaces memcache > directly--simple get/put of key/value pairs--but it also provides atomic list > operations and therefore replaces beanstalk (the queuing system) as well. It > was trivial to rewrite my own job queue system on redis. > > I also took the opportunity to remove direct references to memcache in our > code base and replace them with a generic KeyValueStore class; that way if > redis doesn't cut the mustard I can switch to database tables with little > trouble
redis is indeed a great way to go. mongo is a viable one too. glad it's working! -- -a -- be kind whenever possible... it is always possible - h.h. the 14th dalai lama _______________________________________________ Bdrg-members mailing list [email protected] http://rubyforge.org/mailman/listinfo/bdrg-members
