RE: AoE INT13h progress

> James Harper wrote:
> >
> > With the latter, it allows the BIOS to assume the geometry that the
disk
> > was partitioned with. Not doing so can get you into all sorts of
> > trouble. But maybe have it as optional rather than default
behaviour...
> >
> 
> Yes, the problem with that is you can get into a situation where the
> geometry is wrong, and it stays wrong.

Been there. Linux and maybe some BIOS's do the
infer-the-geometry-from-the-partition-table trick. I had two identical
drives that Linux kept seeing as different geometries. Eventually wiped
the partition table and then rebooted and both were then seen the same.

It's not a huge concern though as geometry isn't nearly as important as
it used to be. It's only really significant in bootstrap.

> > AoE is designed as a SAN protocol, and it is conceivable that no
servers
> > actually have local disks, and that a DHCP/TFTP server might not be
> > available. Also, it would be another point of failure unless the SAN
> > could provide this information... hmmm... definitely warrants more
> > discussion.
> 
> I'm fairly familiar with AoE, and yes, discovery in these sort of
> environments are very much usage-dependent.  Some people will really
> want it, and some can't even handle it :-/

I'm sure we can build something that accommodates several different
discovery mechanisms...

James


-------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc. Do you grep through log files
for problems?  Stop!  Download the new AJAX search engine that makes
searching your log files as easy as surfing the  web.  DOWNLOAD SPLUNK!
http://ads.osdn.com/?ad_idv37&alloc_id865&op=click