On Wed, Oct 08, 2003 at 05:45:34PM +0800, Joon wrote:
> Hi all,
> 
> I really dunno if this really is clustering, but say I have 2 servers
> which I will be using for redundancy (server A goes down, server B takes
> over).  This means I should have exactly the same data on both servers,
> updated in realtime or near realtime.
> 

Yes, this is a high-availability cluster, in active-passive mode.  Yes,
you need to have exactly the same data on both servers, either using
some form of synchronization (a very difficult task, especially if the
data changes often) or use of network attached storage or a storage
attached network (very expensive solution, and probably only appropriate
for high-end database apps).

What application are you interested in?

> Right now I'm reading on Linux Virtual Server.  It seems a viable
> solution, except that I only have 2 machines-- I don't have a director. 
> Any alternatives?

If all you want to do is go active/passive, use Linux-HA heartbeat and
run Apache simultaneously on both servers--all of the time, and ensure
their configurations are perfectly identical.  The only thing in your
resource group is the cluster IP address you want your clients to be
using.  If one server goes down, the IP address gets taken over, back
and forth.

If you only have a web application, and are more interested in load
balancing than redundant failover, your only possible option is to set
up a round robin DNS.  It's by no means as good as LVS, of course, but
those are the breaks.

--
Philippine Linux Users' Group (PLUG) Mailing List
[EMAIL PROTECTED] (#PLUG @ irc.free.net.ph)
Official Website: http://plug.linux.org.ph
Searchable Archives: http://marc.free.net.ph
.
To leave, go to http://lists.q-linux.com/mailman/listinfo/plug
.
Are you a Linux newbie? To join the newbie list, go to
http://lists.q-linux.com/mailman/listinfo/ph-linux-newbie

Reply via email to