7604 SPA interface doesn't support keepalive - how does it monitor?

[jkeeffe]

We have several SPA gig ethernet interfaces that connect to our provider's metropolitan optical ethernet (MOE). Our interfaces connect to the provider's 3750 switch on our premises and the 3750 then connects to their MOE. If the MOE has a problem, the SPA interface goes into line-up protocol-down mode and the EIGRP neighbors go away.

The problem is when the provider fixes the MOE, the link doesn't change from up/down to up/up unless we do a shutdown/no shut on our interface. Then I notice that the SPA interfaces do not support keepalives.

ROC-RT7604A-CR#sh int g2/0/0

GigabitEthernet2/0/0 is up, line protocol is up

Hardware is GigEther SPA, address is 0017.0fac.a600 (bia 0017.0fac.a600)

Description: LLT-RT7604A G2/0/0

Internet address is 164.72.188.6/30

MTU 1500 bytes, BW 1000000 Kbit, DLY 10 usec,

reliability 255/255, txload 22/255, rxload 2/255

Encapsulation ARPA, loopback not set

Keepalive not supported

Full Duplex, 1000Mbps, link type is auto, media type is LX

output flow-control is unsupported, input flow-control is unsupported

ARP type: ARPA, ARP Timeout 04:00:00

Last input 00:00:00, output 00:00:00, output hang never

Last clearing of "show interface" counters 32w4d

Input queue: 0/75/20/0 (size/max/drops/flushes); Total output drops: 0

Queueing strategy: Class-based queueing

Output queue: 0/40 (size/max)

30 second input rate 10617000 bits/sec, 12223 packets/sec

30 second output rate 89996000 bits/sec, 9604 packets/sec

177200803477 packets input, 17873113082287 bytes, 0 no buffer

Received 169 broadcasts (4882913 IP multicasts)

0 runts, 0 giants, 0 throttles

20 input errors, 0 CRC, 20 frame, 0 overrun, 0 ignored

0 watchdog, 6411301 multicast, 0 pause input

20252772537 packets output, 8978256614446 bytes, 0 underruns

0 output errors, 0 collisions, 4 interface resets

0 babbles, 0 late collision, 0 deferred

0 lost carrier, 0 no carrier, 0 PAUSE output

0 output buffer failures, 0 output buffers swapped out

Is that why the interface doesn't transition to up/up when the MOE is fixed because there is no keepalive being generated.

Or is there something else I need to look at to find out why the interface will not automatically come back up?

Remember, we never lose the connection between our SPA interface and the provider's 3750. It is only in the MOE that the connection is lost causing EIGRP to time out and the interface to go UP/DOWN.

[Giuseppe Larosa]

Hello Jim,

the C3750 probably tears down its port when the l2 transport service fails like a CSU on a Frame-Relay service.

Otherwise being your SPA GE connected to a switch port it should stay up/up even when the pseudowire fails and you should detect problems from EIGRP timers expiration.

So the C3750 can be trying to implement some form of ethernet LMI

The missing of keepalive support can play a key role here.

see

http://www.cisco.com/en/US/docs/switches/metro/catalyst3750m/software/release/12.2_46_se/configuration/guide/swoam.html#wp1109491

ethernet OAM

so the state of local port can reflect the result of remote OAM loopback (a sort of end-to-end test)

in your case you could ask to your provider to disable the feature on your side so your interface stays up/up and you use EIGRP to detect faults and to recover.

It's not clear what causes the problem but if there is this workaround available I would try it

Hope to help

Giuseppe

[jkeeffe]

Hi Giuseppe - Our interface does not have ethernet OAM enabled. Should I still contact the provider to see if they have it enabled? If they do, and we don't, could that be causing the up/down state?

[Giuseppe Larosa]

Hello Jim,

the link I've provided is a feature of C3750 ME that I've linked to explain why the port goes down.

I wasn't meaning you should enable ethernet OAM.

I would ask to the provider if they can remove the mechanism that tears down the link when the L2 pseudowire is not working.

Hope to help

Giuseppe