Redundant NFS storage setup (part 2)

Kuba Ober kuba at mareimbrium.org
Wed Nov 28 08:42:37 CST 2007


On Wednesday 28 November 2007, Matthias Saou wrote:
> Kuba Ober wrote :
> > On Friday 23 November 2007, Matthias Saou wrote:
> > > Harry Roberts wrote :
> > > > ReiserFS has supported 16TB filesystems on 32 & 64bit Linux for some
> > > > time, and SGI's XFS supports upto 9PB and aparently should scale much
> > > > better than ext3; I'm not sure how XFS compares to ReiserFS on large
> > > > systems though.
> > >
> > > I've used XFS and ReiserFS in the past, IIRC when ext3 didn't yet
> > > exist, but since ext3 does what I want, and that it's officially
> > > supported in RHEL5, which is what I'll be using, I prefer sticking to
> > > that for now.
> >
> > Before you commit yourself to it, I suggest you create the huge
> > filesystem you think of, use some tool to populate it with non-zero files
> > (a size mix you'd expect in your application), and run fsck. You may
> > change your mind. I've had zero xfs-related problems for several years
> > now, probably dating back to earliest alpha xfs in pre-fedora times, when
> > it wasn't even in the kernel by default.
> >
> > Right now I use it on RHEL4 and RHEL5 for 2 years w/o a single problem.
>
> I'll definitely be considering that, as I already had to wait hours for
> fsck to run on some 2 to 3TB ext3 filesystems after crashes. I know it
> can be disabled, but I do feel better forcing a complete check after a
> system crash, especially if the filesystem had been mounted for very
> long, like a year or so, and heavily used.

The decision process for using ext3 on large volumes is simple:

Can you accept downtimes measured in hours (or days) due to fsck? No - don't 
use ext3.

There's no workaround for that. Do not ever ignore the need to run fsck 
periodically. It's a safe thing to do. You can remount the xfs volume as 
read-only and then run fsck on that - that's another thing to take into 
account when setting things up. If you can partition your application such 
that a large volume that can be made RO, it's useful as it may avoid 
downtime -- if xfs_check reports no problems, there's no reason to totally 
unmount.

Cheers, Kuba



More information about the Linux-PowerEdge mailing list