Re: [ossig] Two obscure questions: directory access times and atomic rename()

[Christopher DeMarco <cdemarco@xxxxxxxxxxx>]

On Tue, Sep 07, 2004 at 02:46:38PM +0800, Dinesh Nair wrote:

> it's  a tradeoff between inodes and   time. they're usually mutually
> exclusive, i.e.  you either  veer  towards saving inodes  or  saving
> time. the middle point is elusive and would differ  from app to app,
> so best to see which profile your  app fits and  then adopt the same
> strategies which others have for mail spools and usenet news spools.

In The Real  World, I've found that the  issue is  actually moot.  I'm
storing stuff in a  tree like the  St. Ives riddle (I  met a  man with
seven wives; each  wife  had seven sacks;   etc.), so each  successive
level of depth in the tree is exponentially-denser than the preceding.
Since I'm storing  image files at ~100k/file,  I have to use trivially
small numbers of inodes (50% subdirectories, 50%  files) or I'm eating
terabytes  of  disk by about my   third or fourth  branch.  The "sweet
spot" seems to be ~36 inodes, where I get 400GB at level four.

So  at this scale the quantum  tunneling effect is unnoticable - shame
'cause I'd like to visit my parents but I can't afford the plane tix.

-- 
% You are in a maze of twisty passages, all alike.
Christopher DeMarco <cdemarco@fastmail.fm>          
PGP public key ID 0x2E76CF5C @ pgp.mit.edu
+6012 232 2106

PGP signature