[Gluster-users] HA strategy

[Keith Freedman]

At 04:51 AM 8/14/2008, Rob S wrote:
>On Thu, Aug 14, 2008 at 11:20 AM, Krishna Srinivas
><krishna at zresearch.com> wrote:
> >>>  My question is: Should I run 2 gluster servers, and several gluster
> >>>clients,
> >
> > Yes that would work nicely.
> >
> >> or could I make each machine a gluster server, and sync using
> >>>AFR?
> >
> > Running server on every client will mean that
> > you will have as many copies as the number of clients.
> >
> > Krishna
> >
>
>Keith, Krishna - thanks for your answers.  I think I've worked it out,
>and wonder if you can confirm my understanding of gluster:
>
>Gluster uses servers and clients.  A client can access a gluster
>resource without requiring the use of any local disk space -
>effectively mounting the remote server locally.  A gluster server
>actually holds the files.  You can have any number of servers and or
>clients.  To achieve ultimate read performance, and where disk space
>isn't a constraint, you could run gluster server on every machine, and
>configure it to read locally.  This would mean that for 1GB of data,
>you'd need 1GB of storage available on every single machine.
>
>Is this right, or have I misunderstood?

that's right.

>Thanks,
>
>
>Rob