There are benchmarks that shows Ext2/3 is slower than all of them.

Ext3 is ext2 with journaling support (basically a transaction log file
called .journal that is stored in the ext2 filesystem).  So there is no
performance gain (except fscking).

I'm using reiserfs on a 2.4.19 kernel with the lastest kernel patches
direct from reiserfs.org.  I'm using it on a 100gb hardware raid array
with 250,000 files and my laptop (the array is not plugged in to my
laptop ;).  I went with reiserfs because of the current speed benchmarks
and future potential.

There is also a difference between meta data journaling and data
journaling.  Meta data is data about the data (file names, file size,
mtime, ctime, atime...).  Journaling meta data ensures integrity of the
file system, which both ext3 and reiserfs do.  Ext3 can journal data (at
a big performance hit), reiserfs does not by design.  They say that data
journaling requires the use of a data-journaling-aware application.  An
application may make two writes as part of one logical transaction.  One
is journaled, then the system crashes.  The file system is intact,
however the data is corrupt or not accurate from an application point of
view.  There is a reiserfs patch to make it do full journaling.  There
is new kernel code for quota support with the reiserfs.  I have not yet
been successful in patching the kernel with the latest reiserfs code and
the quota patches.  It is too much new code for it to make it in to the
2.4 release supposedly.

I know of data corruption in old 2.4 kernels with software raid and
reiserfs.  This was due to the two pieces of code writing on each other.
This has been fixed for some time.

Cory

On Sat, Oct 05, 2002 at 08:10:03PM -0700, [EMAIL PROTECTED] wrote:
> On Sat, Oct 05, 2002 at 06:12:44PM -0700, Bob Miller wrote:
> 
> > > Is there a speed advantage?  I know there's a filesystem advantage
> > > for it if your computer crashes or loses power.
> > 
> > I don't think there's any significant performance change.  Not that
> > I've been able to notice.
> 
> I was reading about ext2 vs. the various journaling filesystem performance 
>benchmarks.  ext2 is faster than any of them.  I've used Reiser and ext3 journalling, 
>and found reiser to be quite a bit faster for my needs.  That said, I use ext3, 
>because I want to have the filesystem protection and some people-- not me-- have had 
>data corruption issues with reiser and the 2.4 kernels.  xfs outperforms either, but 
>*lots* of people have had to deal with hosed filesystems with it.
> 
> Cheers,
> Dennis
> _______________________________________________
> Eug-LUG mailing list
> [EMAIL PROTECTED]
> http://mailman.efn.org/cgi-bin/listinfo/eug-lug
_______________________________________________
Eug-LUG mailing list
[EMAIL PROTECTED]
http://mailman.efn.org/cgi-bin/listinfo/eug-lug

Reply via email to