[Top] [All Lists]

Re: kernel spam when mounting xfs

Subject: Re: kernel spam when mounting xfs
From: Peter Wächtler <pwaechtler@xxxxxxxxxxxxx>
Date: Wed, 05 Sep 2001 17:49:40 +0200
Cc: Linux XFS Mailing List <linux-xfs@xxxxxxxxxxx>
Organization: LOEWE. Hannover
References: <20010904162714.A1567@xxxxxxxxxxxxxxx> <Pine.LNX.4.33.0109052045580.16091-100000@xxxxxxxxxxxxxxxxxxxxxxxxx> <20010905170427.A20994@xxxxxxxxxxxxxxxxxxx>
Sender: owner-linux-xfs@xxxxxxxxxxx
Andi Kleen wrote:
> On Wed, Sep 05, 2001 at 08:53:19PM +0800, Federico Sevilla III wrote:
> > XFS is stable with NFS. I've done relatively small stress tests and know
> > that for loads beyond our typical here, XFS+NFS is stable. Others like Dan
> > Yocum, who is also on the list, I believe have done even more XFS+NFS
> > stress testing. ReiserFS is supposed to be approaching stability with NFS,
> > but ...
> Just to stop the FUD a bit:
> The last known reiserfs NFS problem (not being able to resolve file handles
> again under heavy load) has been fixed with 2.4.5. Before that
> it has been several years been documented as being fixable with a patch.
> The basic problem BTW was that Linux cannot handle 64bit inode numbers.
> XFS runs into the same problem when you start using filesystems >2TB.
Another topic that scares me: Is gcc 2.95.[23] considered bad to compile xfs?
Why is egcs-2.91-something recommended?

I use 2.4.3-xfs on an i686 SMP machine as NFS server, and I had  at least
two hard lockups (the last happened when mounting a CDROM on aic7xxx).
Now if I want to give 2.4.9 a try: do I have to use egcs-2.91 (not there on 

<Prev in Thread] Current Thread [Next in Thread>