|
> I have two boxes I'd like to have setup this way:
>
> 1st Server: Primary Squid and secondary LVS
> 2nd Server: Primary LVS and secondary Squid
>
>
> Traffic from client would be: Client---Squid---LVS---3 backend web
> servers.
>
> If the Squid server went down the primary LVS server would assume the
> role of both squid and the LVS. If the LVS Server went down the
primary
> Squid server would assume the role of the LVS and the Squid server.
> Can this be done?
>
I don't know that this setup works well. For example, if one server (the
frontend) is down, then Squid and LVS run on the same machine, there is
a
problem that Squid cannot access the load balanced service by LVS at the
same machine.
I would like to suggest that you can LVS on one of the backend web
servers, and there are several squid servers before them. Then, you will
have load balancing squid cluster and load balancing web cluster. LVS
run
on one of squid servers too, but LVS handles traffic before squid, LVS
is
used to load balance squid servers.
Thanks for the help but I can't LVS the backend they are
IIS web servers. I want to reduce the amount of servers
used. I want redundancy at all levels. How would I go about
doing this? I need redundancy for the LVS, Squid and web
servers. Would I need two servers for the LVS to load
balance two squid servers then two servers for the LVS to
load balance the IIS servers.
> Both boxes have a default install of Red Hat 9. Does the kernel need
to
> be patched for Red Hat 9. Kinda confused as some places I read it is
> included in 9.
>
I'd like to suggest you build the kernel 2.4.28 rpm with ipvs support
for
Red Hat 9.
Should this not be do difficult based on the how to
docs?
|
