On Fri, 21 Dec 2001, Gabe E. Nydick wrote:
> Will ext3 and xfs live happily together in the kernel? I do not plan to
> run them simultaneously, except while migrating from xfs to ext3. Also, if
> that activity shocks any of you and you have a good reason for me to stick
> with xfs after it trashed two file systems in one day, let me know. Also,
> if there are any unbiased ext3 advocates out there, let me know how
> production worthy you think ext3 is.
I can't comment on ext3, but I might as well get into the
mix on XFS, if for no other reason than to give something back...
We have 3.6 terabytes of XFS on 3 raid-5 systems here, and
a few more installed at client sites, all on Intel systems. I
run it at home on Intel and PPC. We had one minor glitch when
we first set it up, related to putting the journal on the same
device as the FS on a raid system; it got fixed partly with the
help of people on the list, and partly just by kernel updating.
The raid systems are in constant production, and we use them
pretty hard. We hit all kinds of kernel/mobo/smp/apic/whatever
issues that most people never see.
The only data loss I've had came not from XFS, but from a
severely overheated IDE drive in my machine at home (that drive,
with the benefit of a fan, is still running fine, and still
We chose XFS for various reasons at the time. I can't say
that ext3 or ReiserFS wouldn't be equally viable now, but I
have found nothing to cause me to complain about XFS.
---- ``Windows: It does that sometimes.'' -mattdm
Chris J. Bednar
Director, Distributed Computing Product Group