Hi all.
Just a quick thank-you for all the points on this. I've been flat out
for a couple of weeks and not had a chance to test various things but
lots of ideas to attack very shortly.

Peter W.

On 18/04/2007, at 5:34 AM, David Binger wrote:

>
> On Apr 17, 2007, at 11:20 AM, David K. Hess wrote:
>
>> I'm following up on an old thread where this was mentioned:
>>
>>> Another method is to run a StorageServer on another machine that
>>> defers to the primary StorageServer as a "master".  This method is
>>> interesting in that the secondary server can actually be used by an
>>> application.  (See "durus -s -h" for MASTERPORT and MASTERHOST
>>> options).
>>
>> I've searched around and haven't been able to find a good
>> explanation for how this works and how it is intended to be
>> architected into a site.
>>
>> Is the point of this feature to basically reduce/optimize network
>> traffic from a second machine with durus clients to a file based
>> storage server on a primary machine? Or is there some more
>> complicated replication (and fault tolerance) of some sort at work?
>
> I think the main purpose of this option was to demonstrate that
> a StorageServer can be made to work with an underlying Storage
> that itself has invalidations.   It opens the door for some
> creativity with respect to site architecture (and Storage),
> but there isn't any specific intended architecture.
>
> The simple plan still works fine here.
>