OSPF routing issue

[bowser]

Hi,

I have a routing problem on our distribution layer that I'd appreciated some help with.

We have 4 distrubution layer3 3750s (12.2 SEE) attached to core 6509 (SXF). Our distribution layer is configured in their own area (2) and the core is configured for dual areas (2 dist& 0 core).

When a distribution switch goes down it causes packet loss across the whole distrubution. It seems to be when both redundant links fail. From the config below can anyone see why it causes campus wide routing problems? Is this a know fault with OSPF..

router ospf 1

log-adjacency-changes

2 stub no-summary

passive-interface default

no passive-interface Port-channel1

no passive-interface Port-channel2

network xxlan.0.0.0 0.255.255.255 area 2

network xxxp2p.xxx.0.0 0.0.255.255 area 2

-----

interface Port-channel1

no switchport

ip address xxxp2p.xxx..yyy.yy 255.255.255.254

ip pim sparse-mode

ip ospf network point-to-point

ip ospf hello-interval 1

ip ospf dead-interval 3

ip ospf mtu-ignore

hold-queue 150 in

All links are setup in the manner, two on each dist. Any suggestions would be appreciated. Helpful posts rated.

[kamlesh.sharma]

"area 2 stub no-summary "

command has to be in 6509 bcos this core switch will be ABR and publish a default route to ur distribution switchs. in ur 3750's will have only "area 2 stub" command

ur hello timer manupulation has to consistence in ur network so that all the router will update to each other at consistence timing just check u configure the timers each and every interface in ospf domain same.

do u have multiple interface connected to core if yes than packet u will have equal cost load balancing and per-destination load balancing.

please rate if helps

HTH

[bowser]

I have both core & dist setup with:

area 2 stub no-summary

Are you saying best practice is to use

'area 2 stub'

On the dist layer?

Timers are consistant across campus. Thx for help.

[kamlesh1181]

area 2 stub no-summary need to configure on ABR coonected to area 0 so that you will have default route pointing to area 0 core.

Just check ur routing table u should have default route pointing to your core 1 and 2 only not any distribution switch 3750. if u have any default route in 3750 pointing to another 3750 distribution switch you need to filter it out but what i will suggest here is to put ur area 2 as stud only don't make it NSSA (not-so-stub area) so that u have all specific route from area 0 as well as default route in 3750's.