From: Joshua Baker-LePain <jlb17(a)duke.edu>
> The canonical reasons I've heard are 1) they don't want to spend the
> money/time/resources to acquire enough XFS expertise to support it at the
> Enterprise level
I could think of 2 guys they could easily snatch away from SGI that could
bring such experience -- pretty much the 2 behind much of the VFS in
kernel 2.6 anyway (so great resources regardless).
> and 2) besides, as of RHEL4 (they claim), XFS doesn't provide anything
> ext3 already provides, so why bother.
Feature-wise, probably not. The VFS in 2.6 brings a lot of former XFS-only
features to _all_ filesystems. But I still see serious size limitations as
well as scalability to Ext3 versus XFS.
> Yes, I've pointed out on official Red Hat mailing lists that 2 is false
> due (at least) to the issue of backing up ACLs (use star they say -- no
> thanks,
Well, Red Hat has shipped Jorg's "star" in RHL8 on-ward to address
this.
> say I), but I got no response to that. And I've got benchmarks
> showing XFS pretty handily beating ext3 on nice new hardware, but I
> don't have much faith that would get any response either.
To me, I really could care less about benchmarks except when real
performance issues arise.
All throughout 2.4, every other filesystem -- Ext3, JFS (from OS/2**),
ReiserFS, etc... required hack after hack after hack. XFS, with its
"core" additions to kernel 2.4 provided _everything_ standard from
day 1 -- Quotas, POSIX EA/ACLs, etc... It was ported directly from
Irix, and was _entirely_ GPL -- which you'd figure that's something
Red Hat likes.
The only bug that it ever ran into, which resulted in XFS 1.1, was
something that bit me on 2 /var filesystems. Otherwise XFS, like
Ext3 but _unlike_ JFS or ReiserFS, benefits from being the _exact_
_same_ organization for 10+ years. Which means not only is its
on-line operation well-trusted, but it's off-line fsck/repair as well
(something that plagues ReiserFS heavily due to design focus).
Because I could care less how a filesystem works when it works.
I want to know what happens when its inconsistent, and a journal
reply won't solve the problem. I trust e2fsck and xfs_repair.
--
Bryan J. Smith mailto:b.j.smith@ieee.org