From: kang lee (kanghlee@gmail.com)
Date: Thu Jul 17 2008 - 02:38:39 ART
I just want to add few more things on Nate's comment.
If you have sup 720-3B above on 6500, I think you can do GLBP to load
balance switch side or you can do with HSRP.
For server NIC failure, you can teaming to two different switches.
For server failure, you can cluster servers or or make failover server. This
typically requires external storage device. Also, you can do physical to
virtual cluster, if you have VI3 infrastructure. However, clustering has
many restrictions, such as many application is not supported by clustering.
So you can clone the server image and keep it as part of your recovery
plan.
On Wed, Jul 16, 2008 at 10:00 PM, Nate Cielieska <ncielieska@gmail.com>
wrote:
> John,
>
> Typical i have seen this done in the following way:
>
> - Servers have two physical NIC's. These NIC's each have a connection to
> primary/secondary 6500
> - The servers should have a fault tolerant piece of software that allows
> them to "team". Teaming involves spoofing a virtual mac or keeping track of
> 2 Mac addresses and on failure making a card active, the first option being
> more predominant
> - HSRP natively on the 6500 handles this swimmingly. Make a standby group
> and configure HSRP on an SVI relating to those servers. If a switch dies,
> the secondary switch SVI takes precedence, and the teamed NIC fails as well
> giving you your second redundant path.
>
> In reality if you just wanted to cover switch failure, you could do it with
> HSRP on SVI's but your server would be vulnerable to failure at that point.
>
> On Wed, Jul 16, 2008 at 5:29 PM, john white <johnwhite2008@gmail.com>
> wrote:
>
> > Thanks gents for your inputs,
> >
> > The problem is that the client does not want to loadbalance between the
> two
> > servers. Instead he would like one to be active while the other is
> passive
> > ready to kick in in case of failure of the primary.
> >
> > Is there a way out in this case?
> >
> >
> > On Wed, Jul 16, 2008 at 4:50 PM, Rich Collins <nilsi2002@gmail.com>
> wrote:
> >
> > > Shouldn't HSRP be able to handle most of your requirements?
> > >
> > > On Wed, Jul 16, 2008 at 11:41 AM, john white <johnwhite2008@gmail.com>
> > > wrote:
> > >
> > >> Hi Gents,
> > >>
> > >> I have a situation here and would like your assistance. I am preparing
> > for
> > >> my ccie lab and this problem is killing me.
> > >>
> > >>
> > >> I have two Cisco 6506 switches which i would like to configure for
> high
> > >> availability (preferably both active at the same time). I also would
> > like
> > >> to
> > >> run OSPF between the two switches
> > >>
> > >> I have servers that will be connecting to these switches. These
> servers
> > >> are
> > >> in pairs ( i.e two servers both providing the same service to the same
> > >> clients and with same IP address)
> > >>
> > >> I plan to connect each server in a pairing to a different 6506 switch
> to
> > >> so
> > >> that in case of failure of one switch or one server, users can still
> > >> access
> > >> this service via the second server.
> > >>
> > >> The problem is ,
> > >>
> > >> If a server fails , how do the switches route / switch to the second
> > >> server
> > >> ( in the pairing)
> > >> How do i connect these servers to the network bearing in mind they
> will
> > be
> > >> accessed with the same IP Address
> > >>
> > >> Please advice
> > >>
> > >>
> > >>
> _______________________________________________________________________
> > >> Subscription information may be found at:
> > >> http://www.groupstudy.com/list/CCIELab.html
> >
> >
> > _______________________________________________________________________
> > Subscription information may be found at:
> > http://www.groupstudy.com/list/CCIELab.html
>
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
This archive was generated by hypermail 2.1.4 : Mon Aug 04 2008 - 06:11:55 ART