On Jan 23, 7:31 am, JC <[email protected]> wrote:
> The main idea is to flush the cache in a smooth way: i.e. you ensure > that after this flush operation, no data older than what you defined > is in the cache but you didn’t go through the empty cache step. > Basically, this can be done with a flush command that only flushes > data that are older than a given parameter. > In the end, you end up with a refreshed cache as with actual flush > feature but without having had to overload your backup DB. Of course, > DB shall support this traffic but this is always better to spare it! The implementation seems simple enough (although adding to the item overhead is a bit controversial), but it's not completely obvious (to me, at least) that there's a large benefit over a plain flush. It seems to be easier to misuse than use correctly. That is, it feels like it could lead to a lot of confusion when the time ends up overlapping due to values newer than the flush timeline being built upon items that are older than the flush timeline.
