Mailing List Archive: Cisco: NSP

HSRp and OSPF metrics

Index | Next | Previous | View Threaded

mark at noc

Nov 13, 2006, 2:33 PM

Post #1 of 8 (1042 views)
Permalink

HSRp and OSPF metrics

I've got two 4948 L3 switches that have several
netblocks split between them, using hsrp.

That is, under normal operating conditions switch A
is the primary gateway for 192.168.1/24 and switch B is
primary gateway for 192.168.2/24.

Both switches use ospf to let the rest of my network
know about these netblocks.

How can I get OSPF to either

a) redistribute only netblocks for which it is currently the
active HSRP box?

or

b) put a different metric on the redistributed routes depending
on whether the box is active or standby for a particular netblock?

Thanks,
-mark
_______________________________________________
cisco-nsp mailing list cisco-nsp [at] puck
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/

bep at whack

Nov 13, 2006, 3:53 PM

Post #2 of 8 (1049 views)
Permalink

Re: HSRp and OSPF metrics [In reply to]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Mark Kent wrote:
> I've got two 4948 L3 switches that have several
> netblocks split between them, using hsrp.
>
> That is, under normal operating conditions switch A
> is the primary gateway for 192.168.1/24 and switch B is
> primary gateway for 192.168.2/24.
>
> Both switches use ospf to let the rest of my network
> know about these netblocks.
>
> How can I get OSPF to either
>
> a) redistribute only netblocks for which it is currently the
> active HSRP box?
>
> or
>
> b) put a different metric on the redistributed routes depending
> on whether the box is active or standby for a particular netblock?
>

Well, you can't affect the advertising via OSPF based on HSRP currently.
So I ask....

First, why are you redist'ing instead of just sourcing in OSPF?

Second, since both switches should have interfaces in both subnets, you
should set the metrics on switch 1 interface to .1 subnet to be better than
switch 2's interface to .1 subnet and vice versa. Now in steady state, you
will have both routes to both subnets from both routers but with one being
preferred over the other. Then depending on the failure mode that causes
HSRP to failover, one router will stop advertising its routes.

There are some other games you could play with setting of metrics via
route-maps if you absolutely have to redistribute. Also, depending on your
IOS version, you might be able to use EEM to do some config changes upon
notification of a change in standby address.

- --
=========
bep

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.4 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFFWQWFE1XcgMgrtyYRAtN7AKD7pBy3LHJspQBk4b1haZ5O8PbYqQCgreGD
nNfbXzD+9kxJ7DyAYs3DrFA=
=TQMi
-----END PGP SIGNATURE-----
_______________________________________________
cisco-nsp mailing list cisco-nsp [at] puck
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/

mark at noc

Nov 13, 2006, 4:10 PM

Post #3 of 8 (1016 views)
Permalink

Re: HSRp and OSPF metrics [In reply to]

>> First, why are you redist'ing instead of just sourcing in OSPF?

This is probably just a terminology issue...
I meant

redistribute connected subnets

Is that what you call "sourcing"?

Thanks,
-mark

_______________________________________________
cisco-nsp mailing list cisco-nsp [at] puck
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/

bep at whack

Nov 13, 2006, 4:13 PM

Post #4 of 8 (995 views)
Permalink

Re: HSRp and OSPF metrics [In reply to]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Mark Kent wrote:
>>> First, why are you redist'ing instead of just sourcing in OSPF?
>
> This is probably just a terminology issue...
> I meant
>
> redistribute connected subnets
>
> Is that what you call "sourcing"?
>

No. That's redistributing and results in OSPF externals. I'm talking
about using network statements to include the networks for interfaces that
match the network/mask statements.

- --
=========
bep

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.4 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFFWQoiE1XcgMgrtyYRAql3AJ4sNMgXZ2uDec/6JHCNL20q0tcruwCg5eDU
h3RVcEdfZFA3HXa/ljzv7Wg=
=2B+0
-----END PGP SIGNATURE-----
_______________________________________________
cisco-nsp mailing list cisco-nsp [at] puck
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/

p.mayers at imperial

Nov 14, 2006, 4:20 AM

Post #5 of 8 (1016 views)
Permalink

Re: HSRp and OSPF metrics [In reply to]

>>
>> redistribute connected subnets
>>
>> Is that what you call "sourcing"?
>>
>
> No. That's redistributing and results in OSPF externals. I'm talking
> about using network statements to include the networks for interfaces that
> match the network/mask statements.

If you don't have any OSPF peers on that network, why would you want to
do that?

Bearing in mind that changes in externs do not cause an SPF recompute
because they are always at the leaves of the tree.
_______________________________________________
cisco-nsp mailing list cisco-nsp [at] puck
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/

bep at whack

Nov 14, 2006, 12:22 PM

Post #6 of 8 (995 views)
Permalink

Re: HSRp and OSPF metrics [In reply to]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Phil Mayers wrote:
>>>
>>> redistribute connected subnets
>>>
>>> Is that what you call "sourcing"?
>>>
>>
>> No. That's redistributing and results in OSPF externals. I'm talking
>> about using network statements to include the networks for interfaces
>> that
>> match the network/mask statements.
>
> If you don't have any OSPF peers on that network, why would you want to
> do that?
>
> Bearing in mind that changes in externs do not cause an SPF recompute
> because they are always at the leaves of the tree.

Because if you source them, they can be included as stubs in the Router
LSA. Redistributing results in multiple Type 4 (Summary ASBR) and Type 5
(External) LSAs. Further, SPF is cheap in terms of CPU cycles. LSA
generation and flooding is expensive and is a network wide issue for all
routers.

I'd recommend reading "Optimal Routing Design" by Russ White, Don Slice,
and Alvaro Retana. http://safari.ciscopress.com/1587051877

- --
=========
bep

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.4 (MingW32)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFFWiWEE1XcgMgrtyYRAoVTAKDOKhNB+kos71vsm3iq28Ax/avdkQCg8ITc
gqEJOCOwjzSivg976kIlvIo=
=K2f8
-----END PGP SIGNATURE-----
_______________________________________________
cisco-nsp mailing list cisco-nsp [at] puck
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/

ray at oneunified

Nov 14, 2006, 12:46 PM

Post #7 of 8 (1024 views)
Permalink

Re: HSRp and OSPF metrics [In reply to]

>
> Phil Mayers wrote:
> >>>
> >>> redistribute connected subnets
> >>>
> >>> Is that what you call "sourcing"?
> >>>
> >>
> >> No. That's redistributing and results in OSPF externals. I'm
> >> talking about using network statements to include the networks for
> >> interfaces that match the network/mask statements.
> >
> > If you don't have any OSPF peers on that network, why would
> you want
> > to do that?
> >
> > Bearing in mind that changes in externs do not cause an SPF
> recompute
> > because they are always at the leaves of the tree.
>
> Because if you source them, they can be included as stubs in
> the Router LSA. Redistributing results in multiple Type 4
> (Summary ASBR) and Type 5
> (External) LSAs. Further, SPF is cheap in terms of CPU
> cycles. LSA generation and flooding is expensive and is a
> network wide issue for all routers.
>
> I'd recommend reading "Optimal Routing Design" by Russ White,
> Don Slice, and Alvaro Retana. http://safari.ciscopress.com/1587051877
>
Also there are a number of documents at http://www.cisco.com/go/srnd that
provide good background on how to deal with this layer2/3 segmentation and
synchronization and timing.

Suggested document names include:
Campus Network Multilayer Architecture and Design Guidelines
High Availability Campus Network Design-Routed Access Layer using EIGRP or
OSPF

--
Scanned for viruses and dangerous content at
http://www.oneunified.net and is believed to be clean.

_______________________________________________
cisco-nsp mailing list cisco-nsp [at] puck
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/

p.mayers at imperial

Nov 14, 2006, 1:46 PM

Post #8 of 8 (1001 views)
Permalink

Re: HSRp and OSPF metrics [In reply to]

>>
>> Bearing in mind that changes in externs do not cause an SPF recompute
>> because they are always at the leaves of the tree.
>
> Because if you source them, they can be included as stubs in the Router
> LSA. Redistributing results in multiple Type 4 (Summary ASBR) and Type 5
> (External) LSAs. Further, SPF is cheap in terms of CPU cycles. LSA
> generation and flooding is expensive and is a network wide issue for all
> routers.
>
> I'd recommend reading "Optimal Routing Design" by Russ White, Don Slice,
> and Alvaro Retana. http://safari.ciscopress.com/1587051877

Interesting. That's more or less the opposite of what I've understood
from other sources in the past. I'll take a look, thanks for the pointer.

Having said that, we plan to move away from OSPF to to iBGP carrying
routes for other reasons.
_______________________________________________
cisco-nsp mailing list cisco-nsp [at] puck
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/

Index | Next | Previous | View Threaded

Interested in having your list archived? Contact Gossamer Threads