Re: distributed filesystems

[John Reuning <john@xxxxxxxxxxxxxxx>]

I read through the lustre docs a couple of weeks ago.  It looks
promising, but I was left wondering if it would provide the rock solid
stability of fully burned-in technologies.  I was burned a couple of
years ago by repeated fs corruption due to buggy software.  I'm not big
on restoring 100s of GB of data from backups.

>From the lustre docs, it was written to serve HPC cluster needs.  Has
anyone on this list tried to it for large web clusters?

Many thanks,

-jrr

On Fri, 2004-05-28 at 02:12, Joerg Rieger wrote:

> take a closer look at lustre [1].
> 
> [1] http://www.lustre.org/