DragonFly users List (threaded) for 2005-09
Re: UFS filesystem size limit
Joerg Sonnenberger wrote:
On Sat, Sep 03, 2005 at 11:46:57AM +0200, Michel Talon wrote:
As far as recovery after a crash is involved, clearly nothing beats
journalling, and i have to say that performancewise, i have the
impression that Linux journalled filesystems do *very* well
compared to FreeBSD, or at least reiserfs shines when managing a lot of
small files like a news spool and xfs for streaming large files, when
BSD UFS performs well in the medium case.
Don't compare experimental filesystems with UFS :-) But seriously, the
structure of JFS and XFS is very different from UFS, e.g. the use of
btrees for almost anything. That makes them more suitable for some
operations, but horrible for others. It depends on how big your
directories are, how big your files are and all that. There is no
perfect filesystem after all. As you said, the nice thing about UFS is
that it performs well in all cases.
Umm. UFS performs well in all cases is a relative statement. According
to todays standards in the commercial environments for data recovery after
crash, it does not do that well.
As I was saying, once Matt and I are *thorough* with the BUF/BIO system
and finishing journaling, we will have a lot more to say about this, in
terms of practicality.