normally squeaksource is backed up by our university

We planned to buy a really fast server but still Squeaksource -> 
squeaksource 2 is behind the schedule.


On 13 avr. 05, at 16:13, John Pierce wrote:

>
> On 4/13/05, Cees de Groot <cg at cdegroot.com> wrote:
>  
>  In all seriousness: SqueakSource is an extremely important resource, 
> but
> right now too unreliable for regular use - if the proxy isn't down, 
> then
> the server itself is overloaded or otherwise unavailable.
>
>  <snip />
>
>
> This probably means that SqueakSource - at least the MC part - needs 
> to be
> replicatable. Something with my P2P code does mostly automagically. The
> end result could be that www.squeaksource.com points to a list of A
> records, each with a SqueakSource instance. DNS round-robin will take 
> care
> of load balancing, the P2P networking code beneath each SqueakSource
>  instance would make sure that what you upload in anyone instance 
> would be
> shortly available somewhere else.
>
> Could that be a workable setup?
>
>  I'd be for it.  For the record, I think the only contibution I have 
> here is I own the DNS and A records for www.squeaksource.com.  I'd be 
> happy to either enter additional A records to other squeak sources 
> should the P2P networking be baked into SqueakSource.
>
>  Also, I also run my own SqueakSource, mostly because I keep my code 
> backed up and not primarily because kilana.unibe.ch or my dang proxy 
> keeps going on the fritz.  I believe the public SqueakSource web site 
> says you are on your own to ensure you have backups, so putting all my 
> eggs in one basket doesn't sound too appealing to me.  Having multiple 
> replicas of SqueakSource could alleviate that concern a bit if I 
> understood the architecture (i.e. some replication technologies have 
> been know to correctly replicate mass deletions too).
>
>  Regards,
>
>  John
>
>
> -- 
> If you can find a path with no obstacles, it probably doesn't lead 
> anywhere. -- Frank A. Clark