[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: VALinux's 2.4.5 beta kernel with Ext3

On Fri, 8 Jun 2001, Andrew Morton wrote:

> Date: Fri, 08 Jun 2001 21:51:37 +1000
> From: Andrew Morton <andrewm uow edu au>
> Reply-To: ext3-users redhat com
> To: ext3-users redhat com
> Subject: Re: VALinux's 2.4.5 beta kernel with Ext3
> "Bryan J. Smith" wrote:
> >
> > Anyone try this yet?
> >
> > ftp://ftp.valinux.com/pub/software/kernel/beta/2.4.5-beta2va3.11/
> >
> The ext3 patch in that kernel is dated May 30.  A lot of stuff
> has been fixed since then.  The safest and sanest place to get
> ext3 patches at this time is http://www.uow.edu.au/~andrewm/linux/ext3/
> I'll be putting drops from CVS there any time anything significant
> has changed.  Running from CVS is *normally* OK, but if you time
> it wrongly you may pick up a patch which is either broken or has
> unwanted debug code in it.  Best watch the gkernel commit mailing
> list to see what state it's in.
> I've just put up 0.0.6 which, since 0.0.5 has:
> - Fixed a crash which was occurring where metadata was getting
>   incorrectly recycled as data
> - Fixed HIGHMEM support.
> It's in fairly good shape.  I'd recommend against using quotas
> at this time.  Ordered data mode has had the most testing.
> Interactions with loop, highmem, reiserfs and ext2 are OK.
> Interaction testing with RAID and LVM has not even commenced
> (reports welcome!)

I can probably help here testing the above hardware raid and lvm, guess we
can do software raid with journaled fs now too eh?

> I'd recommend building with JBD debugging and buffer tracing
> turned off - there have been more SMP races in the debug
> code than anywhere else, and the buffer tracing code uses
> heaps of memory.  If you *do* get a repeating failure, please
> turn them on again to get diagnostics.

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]