Feasability of VERY large ext3 file system?

Eric Swenson eric at reaction-eng.com
Thu Dec 5 18:57:01 CST 2002

On Thursday 05 December 2002 04:23 am, Arjan van de Ven wrote:
> On Thu, Dec 05, 2002 at 12:20:43PM +0100, Marc Schmitt wrote:
> >
> > Supporting 1TB per filesystem means: "is known to work". Being able to
> > mount a 2TB FS does not mean anything about if it will work or not once
> > you have data > 1TB on it. That's how I understand Arjan's statement.
> More or less correct. I tried to say "1Tb works. 2Tb works or doesn't,
> it really depends on what components are involved.", eg we KNOW lvm and
> md raid will cause problems...

I think my server went down so many times that day that my brain had a fart- 
of course using a 2TB filesystem is different from mounting it.  I agree with 
you on using md and lvm- they cause way too many problems (I did an md5 
config once on an old server... <shudder>).  Good thing I've got that bcm5700 
driver loaded now.  Now if I can only get SGI's replacement redhat kernel to 
compile with the aacraid module working...

Thanks for the sanity checks....


Eric J. Swenson
Information Systems Administrator                eric at reaction-eng.com
Reaction Engineering International               Phone: (801) 364-6925 ext. 26
77 West 200 South, Suite 210                     Fax: (801) 364-6977
Salt Lake City, UT 84101                            www.reaction-eng.com

More information about the Linux-PowerEdge mailing list