[tech] A question of File System

David Luyer david at luyer.net
Sun Apr 21 22:16:58 WST 2002


> The question I have now been asking myself, while having the occasionaly
> random argument with others; is which file system is better: ext3 or
> ReiserFS

"Better"?  There's no such thing as a best filesystem.

Faster for reads?  Faster for writes?  More robust?  More flexible?
Able to grow, shrink and migrate on demand?  With little pink flowers
on top?

Good for reads is NTFS (even though you excluded it as a Microsoft fs...)
or HFS+ I believe also.

Fast for writes... reiserfs if you have small files which grow, ext2 and
ext3 are also pretty fast with a small number of files per directory.

More robust... I'd expect XFS will do well... something which doesn't
need a fsck is good (although early advfs did not have a fsck but was
rather bad because if it ever did get corrupt, the recovery utilities
would typically lock up the system - fortunately after printing an error
about the inode it was trying to fix, allowing you to do a clear inode
on next reboot and start over...).

On demand growth, shrinking and migration... on Digital Unix (aka
OSF/1 or Tru64), advfs rocks for these reasons.  It's really, really
cool.  (PolyCentre advfs that is.)

And if you want pretty pink flowers, well, I'm sure Apple will have
a filesystem with them soon... in five fruity flavours.

> which one is more |337?

ext2fs+inum:patch... (I assume my kernel isn't still running on enyo
with that Grahame? ;-))

or of course there's other kernel hacks like my

/var/spool/mail/x -> /var/spool/.m/x/x

....cuz to be 'leet your fs needs kernel hacks.

Or perlfs.  That's really 'leet :)

> Of course, feel free to suggest a better filesystem still (noone may
> suggest NTFS or FAT*)

Why - you assume Microsoft couldn't make a good fs?  NTFS is actually a
decent fs.

David.



More information about the tech mailing list