OSPF pb

[feberle]

Hi,

I have 2 routers 7513 connected via a satellite link (HSSI interface with QoS configure on it) and some times (it is appenning very ramdomly) the link is falling down during 2 mn.

Here are the conf and logs :

router #1

interface Hssi8/0/0

description *** Liaison Satellite ***

bandwidth 12000

ip address 192.168.201.2 255.255.255.248

max-reserved-bandwidth 80

service-policy output POL2

serial restart-delay 0

Nov 6 15:40:56.537: %OSPF-5-ADJCHG: Process 1, Nbr 192.168.211.26 on Hssi8/0/0 from FULL to DOWN, Neighbor Down: Dead timer expired

Nov 6 15:41:28.180: %LINEPROTO-5-UPDOWN: Line protocol on Interface Hssi8/0/0, changed state to down

Nov 6 15:41:28.248: %STANDBY-6-STATECHANGE: FastEthernet4/0/0.1 Group 0 state Active -> Speak

Nov 6 15:42:28.026: %STANDBY-6-STATECHANGE: FastEthernet4/0/0.1 Group 0 state Standby -> Active

Nov 6 15:42:28.178: %LINEPROTO-5-UPDOWN: Line protocol on Interface Hssi8/0/0, changed state to up

Nov 6 15:42:38.086: %OSPF-5-ADJCHG: Process 1, Nbr 192.168.211.26 on Hssi8/0/0 from LOADING to FULL, Loading Done

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

router #2

interface Hssi8/1/0

description *** Liaison satellite ***

bandwidth 12000

ip address 192.168.201.1 255.255.255.248

max-reserved-bandwidth 80

service-policy output POL2

serial restart-delay 0

Nov 6 15:42:07.178: %OSPF-5-ADJCHG: Process 1, Nbr 192.168.211.25 on Hssi8/1/0 from INIT to DOWN, Neighbor Down: Dead timer expired

Nov 6 15:42:11.602: %LINEPROTO-5-UPDOWN: Line protocol on Interface Hssi8/1/0, changed state to down

Nov 6 15:42:13.606: %LINEPROTO-5-UPDOWN: Line protocol on Interface Hssi8/1/0, changed state to up

Nov 6 15:42:38.346: %OSPF-5-ADJCHG: Process 1, Nbr 192.168.211.25 on Hssi8/1/0 from LOADING to FULL, Loading Done

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

Thks for ur help

[Harold Ritter]

This is not an OSPF problem. The link is going down. OSPF is just reacting to it.

[markneil]

My understanding is that satellite connections can experience delay.

Based upon your logging, suggest that the interval between ospf hello packets may be causing your outages. Suggest that adjusting the ospf dead interval from its default of four times the hello interval (default 10 seconds) may remedy your situation.

ip ospf dead-interval 50

Good luck!

[Harold Ritter]

Changing the OSPF dead timer won't change anything by itself since the adjacency will be reset as the link is declared down:

Nov 6 15:42:11.602: %LINEPROTO-5-UPDOWN: Line protocol on Interface Hssi8/1/0, changed state to down

HDLC will wait 40 seconds before it considers the interface down if it doesn't receive keepalives.

Tell me if I'm wrong but the normal delay for a satellite link is about ~250 ms.

[Harold Ritter]

Correction. A satellite link propagation delay is ~250ms up and ~250ms down = ~500ms total. The point is the problem is unlikely related to the propagation delay.

[feberle]

you're right. The propagation delay (ICMP request) is around 520 ms depending on the data rate and the satellite equipement.

I've found this bug (cscdv28293) about QoS and HSSI but my release (12.2 10.6)is not concerned :

"Add service-policy output to HSSI interface may cause the interface down.There is no workaround."

thks again for ur help

flavien