As the title says, does the ARC cache big files (by "big" I mean 100-200 MB) or does it prefer smaller files and if there's any memory left, cache one or two of the big ones?
Or is it just a matter of number of accesses, that is, files read the same number of times have the same priority to be included in the ARC, irrespective of their size?
Or is it just a matter of number of accesses, that is, files read the same number of times have the same priority to be included in the ARC, irrespective of their size?