One NetApp (regardless of how huge and regardless of how expensive)
means one point of failure, does it not?
Regards.
On Sat, Jan 22, 2000 at 12:05:48PM +0100, Guan Yang wrote:
> An crazy idea for infinite scalability and n-1 redundancy:
>
> 1. A rackful of SMTP servers, running mini-qmail and qmail-qmqpc, behind
> pickdns.
>
> 2. A small amount of qmail servers running qmail and qmail-qpqpd.
>
> 3. A huge, expensive NetApp.
>
> 4. A rackful of POP3 servers, running only qmail-pop3d, behind pickdns.
>
> 5. A rackful of SqWebMail servers, behind pickdns.
>
> 6. A small amount of replicated MySQL servers running authentication.
>
>
> Are there any problems with such an arrangement? Is it scalable? Is it
> overkill for ~1 million users?
> --
> Ideas don't stay in some minds very long because they don't like
> solitary confinement.
