Re: [tlug] how to tune reiser4 for millions of files?

[Francois Cartegnie <fcartegnie@example.com>]

Le jeudi 28 janvier 2010, Michal Hajek a écrit :
> 
> My attention is more on the hw or system side of the problem. That is,
> can I do something with the system (OS, hw..etc.) to speed things up?
Hi,

As Curt told you, you're looking for inode attributes but the file indexing 
hash is just based on names.
Reiser will be fast accessing a file in a large directory assuming you already 
know that filename (ie the hash). Sequentially browsing all files will always 
be slow (and slowest with that B-tree than just a list).

If you're looking for a small subset of inodes or want to only retrieve their 
attributes, give a try to the highly experimental (anyone said as much as 
reiser4?) Mysqlfs. 
This is a really slow filesystem but as all inodes properties will be stored 
in the database it will allow inodes attributes side querying. (and index 
optimizations as well)

Francois