On Sep 20, 12:35 pm, Nbd <[email protected]> wrote:
> On all examples I could find of using memcached it is used as query > result cache. This means that all data manipulations and updates are > done on the db level and not in memory. This is not a recommended usage. Stuff comes out of the database in result set form. I have trouble imagining a case where you don't convert that data into something else (an object or list of objects (for which multiple queries may be required), or in the worst cases, a direct web page). That "something else" is what you should be caching. > once in 10 minutes write all updated users data to database. From where? Hopefully not the cache. Caches are never guaranteed to have data. And you'd still need to know which records, exactly you need updated. It sounds like what you're looking for is a job queue. I'd recommend gearman or beanstalkd depending on your needs.
