[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: JFS / Unsupported file systems



-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512

Alex Samad wrote:
| On Tue, Apr 15, 2008 at 09:49:44PM -0400, Chris Walters wrote:
|> Alex Samad wrote:
|> | Hi
| [snip]
<snip squared>
|> Anyhow, JFS hasn't given me any problems, and at one time I used it for
|> everything except for my /boot partition.
| yes i have read about hans, also read bad things about reiserfs. also
| read bad things about xfs but i have a ups sho that should be okay.
|
| most of the followup googling I did on jfs seems to give the impression
| that jfs is not that fast a fs on lots of small files.

Hi Alex,

Well, every file system has its good points and its flaws.  Reiser set out to
create the "perfect" file system, and ReiserFS performs admirably on lots of
small files - better than any of the others.  Its main flaw is that version 3
is prone to FS corruption, and version 4 is not complete (you can run it, with
a kernel patch, but it is not everything that it was promised to be).

JFS, in my experience performs equally well with all sizes of files, with the
drawback that, with time, it will slow down (probably due to file system
fragmentation).  I find that doing a full backup, running 'mkfs.jfs' on the
partition, and restoring the data fixes that problem.

XFS.  I've used it and, believe it or not, I found it not really much faster
than the EXT3 file system.  Then there is the problem that it uses extensive
buffering to achieve this "speed boost", which I never saw - basically meaning
that you must have a UPS or you *will* lose data.

EXT3.  Of course we all know the benefits and flaws of this one, and its
predecessor.  I thought I would bring it up because it is the 'tried and true'
file system of Linux based systems.

Regards,
Chris
-----BEGIN PGP SIGNATURE-----
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=eFGm
-----END PGP SIGNATURE-----


Reply to: