Hi All,
In my production network, I got kinda the same configuration, but only one
6509 with dual supervisor engines and dual MSFCs configured. The thing is
that this 6509 also run EIGRP and NAT on its MSFCs and things start to get
complicated. Please help me out with this configuration:
- On the first MSFC, all VLANs have higher priority and therefore should be
in active mode.
- On the second MSFC, all VLANs have lower priority and therefore in the
standby mode.
- Both MSFCs have the same static NAT and static routes, default route.
- Both of them have EIGRP with same configuration and have default route
redistributed into EIGRP
Since I started configuring EIGRP on both of these cards, duplicating the
NAT, static routes and other global command on the second MSFC so that it
can be redundant to the first, something started to happen. It seems that:
1. It seems both MSFCs do the routing, even the second MSFC supposes to be
in standby mode.
2. EIGRP does load balancing on both card (I can see multi paths to the
default route
3. Some IP are NATed to the "ip nat outside" VLAN on the first MSFC, while
the others NATed to that of the second MSFC (standby one). Anh some IP are
NATed, but didn't work well because it experiences packet dropping...
Would someone please help me out with this scenario? What did I configure
wrong? Is there any way to keep the second MSFC does NOTHING (eigrp, nat,
static....) beside listening to the active MSFC, and kick off only if the
first MSFC dies... Thanks All!!!!
Thomas N.
""Tony Medeiros"" wrote in message
news:[EMAIL PROTECTED]...
> Found my answer,
>
> From:
>
http://www.cisco.com/univercd/cc/td/doc/product/lan/cat6000/sft_6_1/configgd
> /redund.htm#43570
>
> Quote:
> "Layer 2 redundancy for the supervisor engines (one active and one in
> standby)-If the active supervisor engine fails (the MSFC installed on it
> will also fail), both Layer 2 and Layer 3 functions roll over to the
> redundant supervisor engine and MSFC combination."
>
> So the way I see it:
> For HSRP redundency for the same vlan. The standby MSFC Should be the
the
> MSFC in the redundent supervisor. This will utilize all the replicated
MLS
> information kept in the PFC's (or CEF information on the PFC2) This will
be
> a faster, cleaner failover than having the standby HSRP router in the
other
> chassis in the event of a supervisor failover. It will also keep the
Layer
> 2 path optimal. If the whole box fails then that is a different story.
The
> "listening" HSRP peers will take a little longer to come up as well as
build
> out their MLS/CEF cache.
>
> Cool, and thanks to Jeff for the pointer.
>
> Tony M.
> #6172
>
>
> ----- Original Message -----
> From: "Tony Medeiros"
> To:
> Sent: Saturday, August 11, 2001 1:34 AM
> Subject: HSRP with 4 MSFC's....Best practice [7:15734]
>
>
> > I have a question that really only experience will answer.
> >
> > I have a customer with a two 6509 core with redundant Sup's and MSFC's
in
> > each
> > box ( a little over kill if you ask me) The setup is a typical one with
> > access layer switch's uplinked to each core boxes with the typical L2
load
> > balancing going on by setting up the root bridges appropriately on the
> core
> > boxes. The active HSRP MSFC's correspond with the root bridges. Pretty
> > vanilla setup. A failure of a whole core box will have the layer 2 STP
and
> > layer 3 HSRP's fail over the other core box. They are running a hybrid
> setup
> > with IOS and CAT OS. So far so good.
> >
> > The redundant MSFC's in each box are set up as the lowest priority's so
> they
> > are just in listening mode.
> >
> > My question is: In the case of a primary Sup failure in one of the
boxes
> and
> > the secondary takes over. Will the associated primary MSFC fail as
well?
> I
> > have never tested this and this is a production network so I obviously
> can't
> > test it there. If the associated MSFC fails with the sup, It make
sense
> to
> > me to make the MSFC in the redundant sup the standby one so the layer 2
> path
> > stays optimal and doesn't have to pass over the inter core trunk to hit
> the
> > router in the other core switch. I know this won't be optimal if the
> whole
> > box fails because it takes slightly longer for a listening HSRP peer to
> > transit to active than it does a standby peer to transit to active. I
> could
> > speed things up be lowering the HSRP hello timers some. If the MSFC
stays
> > alive after a sup failure then I think I'll just leave things alone.
> >
> > Has anybody any experience in this type of setup? I didn't build this
> > network
> > but I am tasked to fix a bunch of screwups. I know this would be a moot
> > point
> > if they were running native IOS cus the 2 MSFC's really act as one.
> >
> > Any experience or comments are very much welcome !!!! I haven't posted
a
> > question in quite a while, but it feels good !! I'll be scanning CCO
too
> in
> > the meantime.
> >
> > Thanks in advance
> > Tony M.
> > #6172
Message Posted at:
http://www.groupstudy.com/form/read.php?f=7&i=15911&t=15734
--------------------------------------------------
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
