Re: Filesystem Benchmarks (2.6.0-test2)

[Federico Sevilla III <jijo@free.net.ph>]

On Thu, Aug 07, 2003 at 07:36:52PM -0500, Steve Lord wrote:
> Grr, test2 still has problems, xfs is pretty stable in 2.4, but not
> always so stable in 2.6. There are more fixes coming in test3.

I agree about XFS being pretty stable in 2.4 (since I use it, and so far
so good). I haven't used 2.6, though, so I was unaware of the
XFS-specific problems. I know how to read "test" though, so I guess
problems in general are to be expected.

BTW, out of curiosity, do you think the pending fixes (to be seen in
test3) have anything to do with the performance bits as reported in the
benchmarks done by Grant Miner? Or perhaps as an expert on filesystems
you have comments with respect to the methods employed (eg: were the
tests done sane/fair)? "No comment" will be fine, too. I understand
you're probably very busy.

> Any anyone running a server in production on 2.6 yet needs their head
> examining.

I agree wholeheartedly. I don't think the Anonymous poster I quoted
mentioning his shift to ext3 from XFS was talking about 2.6, though.
That bit seemed to be about XFS in general (eg: 2.4). No mention of what
version of XFS and the Linux kernel that poster used last, though. From
discussions on this list I know that matters significantly.

Maybe I should post on KernelTrap to ask more about the stability
problems the Anonymous poster experienced.

I forgot to mention a more reassuring bit (except perhaps for the last
paragraph) posted by yet another Anonymous poster in the same KernelTrap
thread. To wit:

    Subject: Notice very short SYNC times on XFS
    Author:  Anonymous
    Date:    Thu, 08/07/2003 - 11:51

    That's because it is a real journaling filesystem. Sorry to break it
    to you all. Try ripping out the disks during your script and then
    seeing which one really need not fsck. I'll put money on JFS and
    XFS. XFS is the best filesystem available for Linux - and that's too
    bad people got behind ext3 because it is convenient. Some will say
    that XFS is hacked into Linux from IRIX - good! Any commercial
    discipline rubbing off into the Linux kernel is a good thing. Ever
    hear about those strange bugs involving laptops and EXT3 corruption?

    I'm also upset that RedHat still refuses to make it easy to get XFS
    running on out of the box installs.

    I'll continue to use FreeBSD wherever possible until committing to
    and the using of Linux becomes more of a meritocracy than a
    popularity contest/Mobocracy penguin logos and think geek shirts
    with penguins and Slashdot fan-boys.

> Thanks for the pointer though.

I don't know if I should say "you're welcome".

I think if there's anything that still lets me read these things
(benchmarks), it's that they're not always consistent and it's hard to
tell which methods are believable and which aren't. In my personal
experience, for example, XFS seems to kick ext3's ass. The presence of
big-time XFS users also gives one something of a warm fuzzy feeling
(going through xfs_users.html is a nice thing every now and then).

The recurring surfacing of stability issues during these supposedly
performance-centric discussions also makes one (lay-person) wonder: is
something wrong with XFS or did this person just do things really wrong
with his setup?

 --> Jijo

-- 
Federico Sevilla III  : http://jijo.free.net.ph      : When we speak of free
Network Administrator : The Leather Collection, Inc. : software we refer to
GnuPG Key ID          : 0x93B746BE                   : freedom, not price.