OSPF network unstable

Unanswered Question
Apr 12th, 2009
User Badges:

Hi all,


I have a c3550 and c7513 connected physically to each other and both are running ospf. One day their adjacency just died for about one minute and then went back up again. Below the logs for the c7513:


%LINEPROTO-5-UPDOWN: Line protocol on Interface FastEtherne

t8/1/1, changed state to down

%OSPF-5-ADJCHG: Process 1, Nbr 172.29.104.14 on FastEtherne

t8/1/1.1 from FULL to DOWN, Neighbor Down: Interface down or detached

%LINEPROTO-5-UPDOWN: Line protocol on Interface FastEtherne

t8/1/1, changed state to up

%STANDBY-6-STATECHANGE: FastEthernet4/1/0.5 Group 105 state

Standby -> Active

%STANDBY-6-STATECHANGE: FastEthernet4/1/0.6 Group 106 state

Standby -> Active

%STANDBY-6-STATECHANGE: FastEthernet4/1/0.4 Group 104 state

Standby -> Active

%STANDBY-6-STATECHANGE: FastEthernet4/1/0.7 Group 107 state

Standby -> Active

%OSPF-5-ADJCHG: Process 1, Nbr x.x.x.x on FastEthern

et4/1/0.6 from FULL to DOWN, Neighbor Down: Dead timer expired

%OSPF-5-ADJCHG: Process 1, Nbr x.x.x.x on FastEthern

et4/1/0.7 from FULL to DOWN, Neighbor Down: Dead timer expired

%OSPF-5-ADJCHG: Process 1, Nbr x.x.x.x on FastEthern

et4/1/0.5 from FULL to DOWN, Neighbor Down: Dead timer expired

%OSPF-5-ADJCHG: Process 1, Nbr x.x.x.x on FastEthern

et4/1/0.3 from FULL to DOWN, Neighbor Down: Dead timer expired

%OSPF-5-ADJCHG: Process 1, Nbr x.x.x.x on FastEthern

et4/1/0.4 from FULL to DOWN, Neighbor Down: Dead timer expired


%OSPF-5-ADJCHG: Process 1, Nbr 172.29.104.14 on FastEtherne

t8/1/1.1 from LOADING to FULL, Loading Done

%OSPF-5-ADJCHG: Process 1, Nbr x.x.x.x on FastEthern

et4/1/0.4 from LOADING to FULL, Loading Done

%OSPF-5-ADJCHG: Process 1, Nbr x.x.x.x on FastEthern

et4/1/0.5 from LOADING to FULL, Loading Done

%OSPF-5-ADJCHG: Process 1, Nbr x.x.x.x on FastEthern

et4/1/0.6 from LOADING to FULL, Loading Done

%OSPF-5-ADJCHG: Process 1, Nbr x.x.x.x on FastEthern

et4/1/0.7 from LOADING to FULL, Loading Done

%OSPF-5-ADJCHG: Process 1, Nbr x.x.x.x on FastEthern

et4/1/0.3 from LOADING to FULL, Loading Done


I am running hsrp on this router in standby, another c7513 (x.x.x.x) is active (maybe i should put passive-interface on these subinterfaces).


It has been operating fine previously and the timers (hello, dead etc) are all in sync. Since they are connected physically to each other, I don't see why the problem occurred.


Appreciate if any of you who have experienced similar problems share with me any potential reasons/fixes for this problem.



  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 0 (0 ratings)
Loading.
Richard Burts Mon, 04/13/2009 - 05:43
User Badges:
  • Super Silver, 17500 points or more
  • Hall of Fame,

    Founding Member

  • Cisco Designated VIP,

    2017 LAN, WAN

Muhammad


The log message indicates that the FastEthernet interface went down and came back up. The fact that the standby status of the subinterfaces changed from standby to active is consistent with the interface going down and coming back up. And the change in the OSPF neighbors is consistent with an interface failure. The suggestion to check the other switch to see if it recorded similar issues is a good place to start.


Without knowing more about your topology and what is connected in the various VLANs (are there layer 3 devices other than the 3550 in some of the VLANs running OSPF?) we are not in a good position to adivse whether making the subinterfaces passive in OSPF would be good or not.


HTH


Rick

Giuseppe Larosa Mon, 04/13/2009 - 11:49
User Badges:
  • Super Silver, 17500 points or more
  • Hall of Fame,

    Founding Member

Hello Muhammad,

the log lines you have attached show that interface fas8/1/1 failed line protocol down and line protocol up.

The OSPF adjacency failed and then recovered.


There are no timestamps: so we cannot say how long after interface fas4/1/0 and its subinterfaces experienced the change of state in HSRP and in OSPF.


IF this happened just after the fas8/1/1 had an up/down my guess is that if the first interface is connected to a switch port that is not configured for STP portfast the change of state caused a topology change and STP had to recalculate.

As a side effect during STP recalculation HSRP and OSPF failed on another router interface connected to the same switch or to other switch in the same campus.


Hope to help

Giuseppe


WabukiSensei Wed, 04/15/2009 - 19:43
User Badges:

Hi all,


Thank you for your replies. I'm trying to find the best time to actually do the recommended implementations and get back to you on my findings. Thanks again!

Actions

This Discussion