On Sun, Jul 11, 2010 at 08:05:59PM +0200, Mayuresh Kathe wrote:
> On Sun 11/07/10 23:05, "Ted Unangst" [email protected] wrote:
> > On Sun, Jul 11, 2010 at 4:22 AM, Mayuresh Kathe <mayur...@ka
> > the.in> wrote:
> > Hello, may I know of limitations on supporting large
> > directories (over 5
> > million files) with small files
> > > (less than 10 KB) under FFS/FFS2?
> > > This is for a research project under AMD x86 with
> > SATA Disk[s].
> > It wouldn't be much of a research project if we told you the answer, would
> > it?
> > Step 4 of the scientific method: Perform experiments.
>
> The project is to do with large number of files stored in a directory, but
> definitely not about
> finding out whether OpenBSD would be in a position to handle that.
> The answer is vital to allow me usage of OpenBSD, else I will probably have to
> move over to some
> commercial Unix, hope you can help. :)
>
>
> The project is research, not finding out whether the research wouldn't yield
> results because the
> filesystem couldn't handle management of 5 million small files. :-)
>
man newfs gives the following tantalizing hints:
-b block-size
The block size of the file system, in bytes. If a disklabel
is available, the default is read from it. Otherwise the
default is 16 KB or eight times the fragment size, whichever
is smaller.
-i bytes This specifies the density of inodes in the file system. The
default is to create an inode for each 8192 bytes of data
space. If fewer inodes are desired, a larger number should
be used; to create more inodes a smaller number should be
given.
the rest is left as an exercise to the reader
