opinion on XFS

[Jim Meyer]

On Wed, 2002-05-08 at 12:07, Jarrod Smith wrote:
> On Tue, 7 May 2002, Donald Becker wrote:
> 
> > The SGI release of "Linux XFS" was a port of the Linux code to XFS, not
> > a XFS implementation for Linux.  The port was done in a way that
> > minimized the XFS changes, instead making substantial changes the Linux
> > kernel code.  While the IRIX FS interface does have good design
> > features, few people outside of SGI saw the benefit of having Linux be
> > an IRIX clone.
> 
> I can certainly understand how someone with a large investment in the
> inner workings of the Linux kernel would feel this way.  

It seems to me that we all have a large investment in the inner workings
of the Linux kernel. The more cleanly it works, the better for us; the
more maintainable and accessible the code, the better for us.
Conversely, if the kernel ever gets overwhelmed by cruft or becomes
either unmaintainable or inaccessible to those developers who're sharing
their labor with us, we all lose a large investment of time as we battle
the inherent instability in such a system, not to mention the time we'll
spend researching which platform to leap to next.

In the end, I'll consider the kernel developers to be the experts as
it's them I want to be pleased with the end product. If they're happy,
I'm happy. Mostly.

> However, from a
> user/administrator standpoint XFS is a wonderful option to have.  

No argument here. I like options. =]

> I would not think about using anything else if it would just make 
> it into the mainstream kernel.

Off-topic semi-rant: We should *always* consider alternatives, if only
to educate ourselves. Not doing so creates catchphrases like "Nobody
ever got fired for buying IBM," and "Microsoft-compatible."

> XFS performs flawlessly
> for us.  Performance, flexibility and stability are all there.  I don't
> have to worry about our filesystems anymore and that's a great feeling.

XFS is indeed lovely; I'm an IRIX fan from way back. And, to be clear,
if and when it's accepted into the mainline kernel tree I'll look it
over to be sure it's behaving well in its new home and, if so, I'll see
where it makes sense to use it. 

In the meantime, it's very important to note that it IS available, it's
just that the price is to maintain your own patched source tree and
compile your own kernels. If you don't want to pay the price of your
time, is it really worth that much to you?

> In any case, I really hope XFS makes it into the next kernel.

Me, too. In the meantime, e3fs is working just peachy. =]

Cheers!

--j
-- 
Jim Meyer, Geek At Large                              purp at wildbrain.com