On Sun, Jun 1, 2014 at 12:49 PM, Kevin Burton <bur...@spinn3r.com> wrote:
> It's possible to set caching to: > > all, keys_only, rows_only, or none > > .. for a given table. > > But we have one table which is MASSIVE and we only need the most recent > 4-8 hours in memory. > > Anything older than that can go to disk as the queries there are very rare. > > … but I don't think cassandra can do this (which is a shame). > Cassandra used to offer tunable sized key and row caches per column family. It was decided that it was too much management complexity for too little benefit. My personal view is that there are clearly cases where one wants to prevent one very hot CF from dominating the key or row cache, but where you do want it to be cached. Whether this is worth the overall complexity is not something I have a strong opinion on, but data caching (as opposed to meta-data caching) within the Java heap seems like an exercise in futility to me generally.. =Rob