Solved: ospf nei down

mahesh18 · ‎12-12-2010

hi ,

here is device logs

00318: Dec 12 06:31:55.468 EST: %MSDP-5-PEER_UPDOWN: Session to peer 10.225.205.12 going down
000319: Dec 12 06:32:07.076 EST: %MSDP-5-PEER_UPDOWN: Session to peer 10.226.203.11 going down
000320: Dec 12 06:32:07.508 EST: %OSPF-5-ADJCHG: Process 3738, Nbr 10.225.205.1 2on GigabitEthernet1/22 from FULL to DOWN, Neighbor Down: Dead timer expired
000321: Dec 12 06:32:41.876 EST: %MSDP-5-PEER_UPDOWN: Session to peer 10.225.205.12 going up
000322: Dec 12 06:32:43.656 EST: %MSDP-5-PEER_UPDOWN: Session to peer 10.226.203.11 going up
000323: Dec 12 06:33:09.612 EST: %PIM-5-NBRCHG: neighbor 10.27.12.21 DOWN on interface GigabitEthernet1/22 (vrf default) DR
000324: Dec 12 06:33:09.612 EST: %PIM-5-DRCHG: DR change from neighbor 10.27.12.21 to 10.27.12.21 on interface GigabitEthernet1/22 (vrf default)

can some tell me if ospf nei us up or down i check the int gi1/22 it is up up

GigabitEthernet1/22 is up, line protocol is up (connected)

also

sh ip ospf int gi1/22
GigabitEthernet1/22 is up, line protocol is up (connected)
Internet Address 10.27.12.21/30, Area 119
Process ID 3738, Router ID 10.27.12.15, Network Type POINT_TO_POINT, Cost: 2
Transmit Delay is 1 sec, State POINT_TO_POINT
Timer intervals configured, Hello 10, Dead 40, Wait 40, Retransmit 5
oob-resync timeout 40
Hello due in 00:00:02
Supports Link-local Signaling (LLS)
Index 10/10, flood queue length 0
Next 0x0(0)/0x0(0)
Last flood scan length is 1, maximum is 40
Last flood scan time is 0 msec, maximum is 12 msec
Neighbor Count is 0, Adjacent neighbor count is 0
Suppress hello for 0 neighbor(s)

thanks

mahesh

lgijssel · ‎12-12-2010

Well, it looks as if this interface has no neighbors:

Neighbor Count is 0, Adjacent neighbor count is 0

Looks as if you are having an issue in a datacenter.

This is generally quite complex and cannot be resolved with such tiny bits of info.

We need the bigger picture to be of help.

regards,

Leo

please remember to rate helpful posts.

View solution in original post

Richard Burts · ‎12-13-2010

Mahesh

It may help to think about the extent of the layer 3 domain (what is the boundary of the layer 3 domain to which your switch is connected), the extent of the layer 2 domain (what is the boundary of the layer 2 domain to which your switch is connected) and the boundary of the layer 1 domain (what is the boundary of the layer 1 domain to which your switch is connected).

The layer 3 domain is the group of devices that are locally connected at layer 3. Since your switch and its OSPF neighbor are in the same subnet they are in the same layer 3 domain, and since you have indicated that this is functionaly a point to point connection they both function as boundary points for the layer 3 domain. The fact that you could not ping to the OSPF neighbor device is an indication that there was a problem within the layer 3 domain.

The layer 2 domain is the group of devices that are locally connected at layer 2. In early times of networking, for example 10Base2 or 10Base5, the boundary of layer 2 and of layer 3 would be the same. But as networking progressed from the use of bridges and hubs, to the use of switches they boundary of layer 2 became different from the boundary of layer 3. It is very likely that your switch has a layer 2 boundary at the provider device that gives you an Ethernet handoff to the MAN network. It is likely that there are layer 2 boundaries within the provider network, and that there is a layer 2 boundary between the provider device and the OSPF neighbor device. Given your post that the provider was doing maintenance we can assume that there was a problem within the layer 2 domain within the provider network, while your layer 2 domain of your device remained up (and without problem).

The layer 1 domain is the group of devices that are locally connected at layer 1. In you case there is almost certainly a layer 1 domain between you and the provider device that gives an Ethernet handoff. And there are other layer 1 domains between the provider devices, and another layer 1 domain between the OSPF neighbor device and its provider Ethernet handoff.

So yes we can say that layer 1 and 2 were down somewhere (within the provider network) but that layer 1 and 2 remained up at your site and at the site of the OSPF neighbor.

HTH

Rick

HTH

Rick

View solution in original post

lgijssel · ‎12-12-2010

Well, it looks as if this interface has no neighbors:

Neighbor Count is 0, Adjacent neighbor count is 0

Looks as if you are having an issue in a datacenter.

This is generally quite complex and cannot be resolved with such tiny bits of info.

We need the bigger picture to be of help.

regards,

Leo

please remember to rate helpful posts.

mahesh18 · ‎12-12-2010

Many thanks for reply

it is point to point man connection between two sites.

routers interface on both sites are up.

do you know what i can check next?

mahesh

lgijssel · ‎12-12-2010

So this is an external connection between sites?

In that case, I should contact the provider because the circuit seems to be failing.

The behavior you describe is typical for ethernet leased lines over a shared infrastructure.

They are terminated on a local switch at customer's site. As long as the switch remains powered, the interface to the customer stays up but there is no connection because the provider side has lost connection with the network or the logical circuit to the other side is interrupted.

You can verfiy this by trying to ping the other side. If this succeeds, the issue is with your equipment. If not (what I would expect to see here) then resort to the advice above and call your provider. If the circuit is owned by your company, contact the people responsible.

regards,

Leo

Please remember to rate helpful posts.

mahesh18 · ‎12-12-2010

thanks for reply.

i am unable to ping the nei ospf from either sites.

it is MAN connection.

so i understanding is layer 1 is up but 2 is down right ?

i will now call the provider

thanks

mahesh

Ganesh Hariharan · ‎12-12-2010

Many thanks for reply

it is point to point man connection between two sites.

routers interface on both sites are up.

do you know what i can check next?

mahesh

Hi Mahesh,

If the neighbour link is up and up state and still neighbourship is going up/down,it can be problem with low memory with router.check out the log messges in routers what it has displayed.

Hope to Help !!

Ganesh.H

lgijssel · ‎12-12-2010

Hi Ganesh,

I tend to disagree on this one because the log shows a dead timer expired.

As we are only seeing one side it may be the ospf process on the other end having problems but that is not obvious.

Proper troubleshooting dictates to check the obvious things first, router memory problems are not in that list.

After all, shouldn't we assume that this config has been working until the problem occurred?

How likely does this make memory issues as a potential cause?

regards,

Leo

mahesh18 · ‎12-12-2010

Here is log from other end router

000115: Dec 12 06:33:09.631 EST: %PIM-5-NBRCHG: neighbor 10.x.x.x. DOWN on interface GigabitEthernet1/0/25 non DR

lgijssel · ‎12-12-2010

Well, I believe it is down but this message is from PIM, implying you have multicast configured here as well.

Now, the ip of this side is: 10.27.12.21/30

The question is: can you still ping neighbor 10.27.12.22?

regards,

Leo

Please remember to rate helpful posts.

mahesh18 · ‎12-12-2010

thanks for reply

i can not ping the nei ip.

i can only ping the local routers ip address?

how u know if multi cast is configured by message --PIM ?

is there any command we can use to know if multicast is configured?

regards

mahesh

lgijssel · ‎12-12-2010

Hi Mahesh,

Please await the results of your call to the provider.

Regarding Multicast: I presume you would know if it is configured.

regards,

Leo

Please remember to rate helpful posts.

Mahesh Gohil · ‎12-12-2010

Hello Mahesh,

mahesh here.

> Regarding PIM:

> To see if multicast is enabled or not you can issue below command

       #sh ip multicast
        Multicast Routing: enabled

> To see PIM enabled interface issue below command

#sh ip pim interface

> Regarding OSPF neighborship

I feel it is to early to watch ospf log as first step should be ping from end to end. It could be L1 or l2 problem because there are logs for

interface up/down. You can probably push provider for this and I am sure they will help to resolve this. From your end you can attach

other device to this port to see if your port is ok or not (to check for any harware problem)

Hope this helps

Regards

Mahesh

mahesh18 · ‎12-12-2010

Hi everyone,

it was telco maintenance.

ospf is up now.

i ping the ospf nei it is pingable

also i did

#sh ip ospf int gi1/22
GigabitEthernet1/22 is up, line protocol is up (connected)
Internet Address 10.x.x.x/30, Area 119
Process ID 3738, Router ID 10.x.x.x., Network Type POINT_TO_POINT, Cost: 2
Transmit Delay is 1 sec, State POINT_TO_POINT
Timer intervals configured, Hello 10, Dead 40, Wait 40, Retransmit 5
oob-resync timeout 40
Hello due in 00:00:05
Supports Link-local Signaling (LLS)
Index 10/10, flood queue length 0
Next 0x0(0)/0x0(0)
Last flood scan length is 1, maximum is 40
Last flood scan time is 0 msec, maximum is 12 msec
Neighbor Count is 1, Adjacent neighbor count is 1*******************

it shows nei count now as 1.

my question is as interface was up so it was not layer issue right?

was ospf nei down due to layer 2 ?

it was MAN connection and as my understanding MAN is layer 2 only and no routing involved?

correct me if i am wrong.

Richard Burts · ‎12-12-2010

Mahesh

I do not fully understand your question, but let me respond to a few parts of it.

You say:"MAN is layer 2 only and no routing involved'. The service provided by the MAN is typically a layer 2 service that provides connectivity from your device to other devices at other locations. The provider is not doing any routing on that connection. But you can do whatever you want to do over that connection including running a routing protocol - as you obviously are doing.

So no routing being done by provider but possibly routing being done by customer.

Part of the confusion in the discussion is that the device interface indicates that line is up and line protocol is up. But that is an indicator of the state of the connection from your device/switch to the next device, which would be some kind of Ethernet handoff from the provider to you at your site. It does not tell anything about the state of the connection to your neighbor. In the case of this discussion the local link was up/up but no connectivity to neighbor because of provider maintenance. In this kind of situation the results of ping are a much better testing tool than the output of show interface.

HTH

Rick

HTH

Rick

mahesh18 · ‎12-12-2010

Thanks Richard for reply.

So our router interface to service provider device was up this means it can be layer 1 or 2 issue from providers devices at both sides.

ping was not working so we can say that layer 1 and 2 were wdown right?