Courtesy of Google...
 
Fragmentation of ReiserFS and some others:
http://www.informatik.uni-frankfurt.de/~loizides/reiserfs/
 
General performance of file systems:
http://www.namesys.com/internal-benchmarks/benchmark-results.html
http://www.namesys.com/whole.shtml
 
Bear in mind that these scores appear to be based on ReiserFS 3.x and earlier, while the new 4.x one will be quite a bit faster (can't find the link at the moment to the graph, but yeah it's an improvement).

there aren't any defraggers on linux as it's unnecessary. the filesystem doesn't fragment like it does in Windows.

http://sdb.suse.de/en/sdb/html/ext2frag.html
http://www.science.unitn.it/~fiorella/guidelinux/tlk/node101.html
 
but if you really want a defragger they are out there as this one points out
http://www.oreillynet.com/pub/faqs/linux_faq_AEN1511
 
essentially it comes down to this... Microsoft opted for speed in the filesystem, shove the file in the first available block etc... On Linux the filesystem is designed to find enough contigious blocks to accomodate the file whenever possible. fragmentation does occur, but it's pretty minimal... not enough to sweat over.
 
By the way, thanks for all your posts here. I've been lurking for a couple years now and have learned quite a bit about tweaking Windows from your posts

http://www.washington.edu/R870/examples/fragmentation
Maybe this will help

Here some info on the ext3 file system:
http://www.redhat.com/support/wpapers/redhat/ext3/
 
The latest version of Defrag seems also to work with ext3 partitions:
http://packages.debian.org/stable/admin/defrag.html
 
From the Debian changes mailing list:

Quote: Changes: defrag (0.73pjm1-1.woody.0) stable; urgency=high
 
* ext2.c: Gracefully fail on filesystems with journals (closes: #118635).
This fixes a bug that causes filesystem corruption when defrag is run on ext3 filesystems.
http://lists.debian.org/debian-changes/2002/debian-changes-200211/msg00016.html
 
Unfortunately, the latest RPM for Red Hat is dated Sep 1 1998:
http://rpmfind.net//linux/RPM/contrib/libc6/i386/defrag-0.73-5.i386.html

Quote: Redhat "phoebe" build 8.1 with the new KDE!

Just FYI: Lot things are broken in the first 8.1 beta release. You might want to check out the Phoebe beta mailing list at https://listman.redhat.com/mailman/listinfo/phoebe-list

Concerning Defrag. I found this on Red Hat's ext3-users mailing list:
https://listman.redhat.com/pipermail/ext3-users/2002-May/003462.html

Well, it's been a couple weeks now since this thread was active. I was wondering how the test went.. is going... will go... come on! Inquiring minds want to know!!
 
seriously tho... any enlightenment on whether this whole ext2 needs no stinkin defraggin thing is a hoax?

I totally see and agree with what you say in your example. the only thing I was wondering was whether or not that scenario occurs enough in the "real world" for defraggers to really be necessary on ext2. In this age of 100+ GB drives do many people ever really fill the drives over say 70% (pulled that number out of my butt). :x
 
I was looking around the net to find people that have asked the same question as you did, only to find one person who had written a program that would create and remove several large and small files just to see if he could fragment the drive. Apparently from his test he was able to fragment it heavily and perfomance did suffer from it. He also asked at the end of his article whether his test was real world or not. Whether he had "overdone it".
 
Anyway... everything I saw effectively said, yea fragmentation does occur... just not enough to really amount to anything worthwile. I myself am curious if this is really true. I don't really do that kind of stuff with my computer and have had Linux and FreeBSD systems run for long (over a year) periods with no noticeable slowdown. I'm not knowledgeable enough to know how to check for it other than when the system boots as it mounts the partitions it tells x.x% non-contigious etc...
 
Regardless, if you ever get around to it and discover anything either way please come back here to let us know. I for one would genuinely like to know
 
Jim