Connectivity problems between Catalyst 6009 and Cisco 7507

[etamara]

Hi,

I experienced something really weird. I have a Cisco 7507 router connected to a Catalyst 6009 by means of a 100BaseTX FasEthernet connection. I powered the switch off as a high availability test ( there is also another Cat6009 connected to another Cisco7507. Cats are directly connected by means of a GigaEtherChannel. HSRP is configured on the 7507s for the unique LAN segment ).

To my surprise, the FastEthernet interface of the neighboring 7507 remained "up, up", even though no neighbor was seen by CDP on that interface!!! As a result, this router had a direct route to the LAN segment, which makes it work as a blackwhole for remote offices whose WAN endpoint is on this router.

Does anybody have a clue about what caused this unexpected behaviour? What could be the reason why the FastEth interface didn't sense the loss of connectivity? What can be done to avoid this problem?

Thanks,

Enrique

[smif101]

I really don't see how that is possible. Did the link light stay green on the router while the switch was powered off? I have never seen an interface without a physical connection stay up. If you disconnect the cable do you still get the same result?

[s.uslay]

and there is no any other media converter/transceiver between 7507 and 6509 ? That could be the only possibility the keep the link up.

The other weird thing could be forgotten loopback termination...

SU

[Endwigast]

what is the default gateway of your core switches? both switches should have the hsrp ip as the default gateway.

[etamara]

Hi,

We actually have two IP subnets mapped to the same (unique) VLAN by means of the "ip secondary" command:

- 180.180.0.0/16 is for user data.

- 180.185.0.0/16 is for managemente traffic.

Switches' interface sc0 is configured with a 180.185.0.0 ip address and the default gateway is the secondary HSRP address ( also 180.185.X.X ).

Enrique

[etamara]

Hi,

There is nothing between the devices. Just the copper cable...

What do you mean with loopback termination?

[s.uslay]

so you have this ?

WAN WAN

7507 (1) ---6509(1) ---- ---6509(2) -------7507 (2)

---------------------

etherchannel

and you shutting down 6509 (1) ??

Are 7507s connected by other means ?

please explain a bit ? Where is HSRP definition ?

SU

[etamara]

so you have this ?

WAN WAN

7507 (1)---6509(1)----EtherChannel---6509(2)-------7507 (2)

and you shutting down 6509 (1) ?? YES

Are 7507s connected by other means ? YES: serial back-to-back

please explain a bit ? Where is HSRP definition ? HSRP is defined in 7507(1) and 7507(2) FastEthernet interfaces ( connecting to 6509s ). I have just one VLAN, which is mapped to two IP networks ( using "IP secondary" ): 180.180.0.0/16 for user data, and 180.185.0.0/16 for SNMP data. Both IP networks have HSRP configured. The active router, under normal conditions is 7507(1). The problem is that HSRP becomes active on both 7500 as a result of the FastEthernet interface remaining up. The key point is why the Fast interface remains up after shutting down 6509(1).

Thanks,

Enrique

[smif101]

If you shut down the 6509(2) switch does the 7507(2) router stay up like the 7507(1)? Have you tried replacing that copper cable or connecting the 7507(1) to another switch and see if the same thing happens, this way you can isolate it to a 6509 or 7507 problem and then go from there. I am going to guess that the 7507 ethernet interface is flaky.

[etamara]

Hi, guys

Problem sorted out! Keepalives were disabled on the interface ( misteriously ). Therefore, the "line protocol" remained in the same state as the administrative status.

Thanks you all for your help and interest!

[showmanz]

I can't belive that! A typical serial line failure involves losing Carrier Detect signal. Because this sort of failure is typically noticed within a few milliseconds