<OT> XFS Build Problems

[Collins]

On Fri, 03 Jan 2003 20:46:55 -0700
Andrew Mathews <andrew_mathews at linux-works.org> wrote:

> Collins wrote:
> <snip>
> > Nor has EXT3 ever failed this (usually unintentional) test for me. 
> > I did have a reiserfs system fail to recover after a lockup (about 3
> > years ago), but I'm user the current reiserfs is stable now.
> > 
> 
>   I truly wish I could say the same, but unfortunately that was one of
>   
> the failings we could demonstrate fairly consistently. During a 2 hour
> 
> evaluation before our Chief Justice, CEO, and CIO, as well as the 
> management team, ext3 never did survive 5 hard resets in a row. We 
> rebuilt the machines exactly the same, only difference being on an xfs
> 
> filesystem, all packages were the same. xfs still, to this day, has 
> never failed to recover itself, even under almost 100% load. We
> migrated our Informix database from raw logical partitions under AIX
> to xfs partitions under linux and the data set was monitored for
> corruption at the moment of impact with absolutely NO loss. 

[ snips ]

>   Ext3 works great for most people, but I've had and seen too many 
> problems to consider it ready for a production machine in our 
> environment. I'm glad you've had good luck with it though.
> 
> -- 

Good to know.  

I'll probably move to XFS when it becomes part of the
mainstream in the 2.6 kernel.  I hate patches, most especially patches
that aren't always compatible with the rest of kernel development. 
Also, in the past, not every distro has XFS support, so I would get into
the catch-22 situation of needing something from an XFS partition and
not being able to mount it.

-- 
Collins Richey - Denver Area
gentoo 1.4 system