Just to add what Chuck said. We have Compaq servers
at work with Compaq NICs. The nice thing about them
is the software allows you to tie several interfaces
together. This is called "Teaming." A single MAC
address is shared between three physical interfaces.
One interface is even Gig Ethernet, and the other two
are fast ethernet. When the GigE fails and looses
link, the NIC driver will send a broadcast on a backup
interface using the source MAC address previously used
on the GigE. This refreshes all CAM entries on the
switches and reroutes traffic to the backup interface.
No IP addresses change. Your customers will never
know anything failed if you design your network
properly.
Chris M.
--- Chuck Larrieu <[EMAIL PROTECTED]> wrote:
> Surprise!
>
> Just had this conversation with a customer the other
> day. They are buying a
> new switch, but want to keep the old one on line to
> be the backup / failover
> for their single NIC servers. How do they do this?
>
> 1) Windoze servers, anyway, require different IP
> subnets on each NIC in the
> server. So now you are into VLAN's and routing on
> your core, whether you
> like it or not. Don't know from experience ( or lack
> ) if this is true in
> *nix boxes. Help?
>
> 2) One thought - assume the switch itself doesn't go
> down, but the blade
> does. Plug the different NICs into different blades
>
> 3) OR - have dual switches, and plug one NIC into
> each switch
>
> 4) Nice thing about IPX - Novell servers can have
> multiple NIC with the SAME
> network address. Makes life simple.
>
> Funny how the issue of redundancy becomes so complex
> so quickly
>
> Chuck
>
> -----Original Message-----
> From: [EMAIL PROTECTED]
> [mailto:[EMAIL PROTECTED]] On Behalf Of Dave
> Ng (Dragon)
> Sent: Monday, December 18, 2000 9:34 AM
> To: [EMAIL PROTECTED]
> Subject: OT-Redundant switches at the servers
>
> I've looked through the BCMSN book and also online
> and I've found info on
> redundant switches and routers at different layers
> in a network (Access,
> Distribution, and Core) but it seems that in every
> case (at least the ones
> I've seen) the redundancy does not flow down to the
> actual physical
> connection to an end node (specifically a server).
>
> Here's the problem I need to solve:
> I have a 3 servers in which I will put redundant
> NICs but do both NICs
> connect to the same switch? Then is that switch a
> single point of failure?
> Is there a way to connect each NIC to a different
> redundant switch and when
> a switch fails the other takes over? Can some one
> point me in the right
> direction? Perhaps a Cisco whitepaper or a book
> that deals with this.
>
> Thanks,
>
> Dave
>
>
> _________________________________
> FAQ, list archives, and subscription info:
> http://www.groupstudy.com/list/cisco.html
> Report misconduct and Nondisclosure violations to
> [EMAIL PROTECTED]
>
> _________________________________
> FAQ, list archives, and subscription info:
> http://www.groupstudy.com/list/cisco.html
> Report misconduct and Nondisclosure violations to
[EMAIL PROTECTED]
__________________________________________________
Do You Yahoo!?
Yahoo! Shopping - Thousands of Stores. Millions of Products.
http://shopping.yahoo.com/
_________________________________
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
