On Sat, Oct 15, 2016 at 6:25 PM, Gordon Messmer <gordon.mess...@gmail.com>

> On 10/13/2016 11:49 PM, c...@zip.com.au wrote:
>> Out of interest, how recent was your need to recover XFS, and what were
>> the circumstances that caused it to be necessary?
> I don't remember the date, specifically, and it's not clear what the
> circumstances are.  One of the consequences of on-line fs checking is that
> errors tend to be detected long after they actually occur.

Not sure what you mean -- certainly with most linux distros fsck gets run
at intervals on clean reboot and does on rare occasions find problems to

I've never seen a problem with XFS filesystems on Scientific Linux 7, but I
have a lot more system hours experience with xfs on SGI Irix, where the
first indication that a disk was failing would come from errors running the
nightly fsr.  When XFS was first ported to linux I used it on 32-bit linux
to recover data disks removed from dead SGI workstations.  The XFS
utilities were extremely slow in that environment.  One reason given for
the poor performance was that XFS was written with deeply nested function
calls with long argument lists that were very inefficient on 32-bit Intel
hardware at that time.

George N. White III <aa...@chebucto.ns.ca>
Head of St. Margarets Bay, Nova Scotia
users mailing list -- users@lists.fedoraproject.org
To unsubscribe send an email to users-le...@lists.fedoraproject.org

Reply via email to