On Thu, Oct 17, 2013 at 01:47:15AM -0400, Richard Yao wrote: > > > On Oct 16, 2013, at 8:13 PM, Saso Kiselkov <[email protected]> wrote: > > > On 10/17/13 12:42 AM, Richard Yao wrote: > >> One of these days, I will learn to include the CC list when using my > >> iPad. Anyway, I like being able to tell people that ZFS is self > >> tuning. :) > > > > Self-tuning is good, but you can't possibly self-tune everything every > > time. I consider running a large dataset on recordsize=512 to be an > > extreme edge case and edge cases are notoriously difficult if not > > impossible to design for in advance. > > It is not necessary to go that far. There do exist servers with many small > files. Mail servers are one example. These servers will tend to have small > ARC buffers.
You know, a dynamically sizing hash table would solve this problem. But we've already tossed that idea out. ;) It's a hash, if the chains are "long" the hash implementation is broken, imo. Using AVL trees to fix a broken hash implementation just seems wrong. -- Cheers, Prakash > > > > Cheers, > > -- > > Saso _______________________________________________ developer mailing list [email protected] http://lists.open-zfs.org/mailman/listinfo/developer
