Terracotta Discussion Forums (LEGACY READ-ONLY ARCHIVE)

There are multiple caches.

I was running refreshes like every 5 minutes. I did have a look through the code and found that it was doing alot more cache writes then i expected when doing updates, i've optimized (dont delete/write if i dont have to etc) that so I will let you know how that goes after testing.

The selfpopulatingcaches don't actually have much in them, but there refreshes update some elements in other caches (pretty convoluted i know it just evolved that way when i went to bigmemory).

So my optimisation has drastically reduced the rate of growth of my diskstore, but it is still growing.

I've also confirmed im not mistakenly actually creating more keys, all im doing is updating existing elements.

Is there a way to determine what policy the compactor is running with?

I worked out how to change the compaction policy, tried SizedBasedCompationPolicy but it just chewed all the disk up quickly.

So I found the offline compactor, when i run it over my cache data it does significantly reduce it's size. So I think we can assume that the compactor threads are never running for whatever reason. Is there any way to progamatically trigger the compaction? Or do I need to roll my own?

I really need to solve this problem. Ideas?