[linux] software RAID-1 query

[Evan Martin]

On 4/25/05, Garrett Cooper <youshi10 at u.washington.edu> wrote:
>      Just thought I'd add to the discussion a bit.
>      I know it's a little off-topic, but just wanted to note the fact that
> differing fs'es get different things done according to the task. Note this
> article with benchmarks on Linux Gazette for some info:
> http://linuxgazette.net/102/piszcz.html.

I think that article was well-intentioned but the things they're
benchmarking are not especially useful.  Doing anything in a directory
with 10,000 entries is benchmarking a specific aspect of the file
system that isn't commonly used.  Try searching for variations on
"ext2 hash trees" for some discussion of the different data structures
potentially usable for these sorts of operations.