OSPF Hello and Dead Interval

This topic is to discuss the following lesson:

If you change the hello and dead time on R1, do you have to make the exact same change on the R2 interface? Do the hello and dead timers have to be the same on both routers?

Hi Jason,

These have to be match yes, otherwise no neighbor adjacency is established.

Rene

Dear Rene,
Could enabling BFD be better option than reducing dead-hello timers regarding the consuming router resource?
Which method do you recommend ?

Onur,
Using Bi-Directional Forwarding Detection is generally considered to be the better option for a couple of reasons:

  1. BFD can detect a downed neighbor much faster than any OSPF timer manipulation
  2. BFD is less CPU intensive than very small OSPF timers because BFD is pushed off to the data plane, whereas OSPF timer events must be processed by the router’s control plane
1 Like

Dear Rene,

Many Thanks to you .One question…

Why Neighbor will not form if Hello/dead Mismatch ?? Could you please explain behind the reason on it ? I think if Hello/Dead is not same on both end then there could be operational related problem like One router release neighbor first while other router still wait, right ??

br//
zaman

Hi Mohammad,

A good question. I am referring to you to RFC which is a type of publication from the Internet Engineering Task Force (IETF) and the Internet Society (ISOC), the principal technical development and standards-setting bodies for the Internet.
On April 1998, they have created the publication RFC 2328 for OSPF version 2, and on point 9.5 they spoke about “Sending Hello Packets” as following:

“Hello packets are sent out each functioning router interface. They are used to discover and maintain neighbor relationships.[6] On broadcast and NBMA networks, Hello Packetsare also used to elect the Designated Router and BackupDesignated Router.The format of an Hello packet is detailed in Section A.3.2. The Hello packet contains the router’s Router Priority (used in choosing the Designated Router), and the interval between Hello Packets sent out the interface (HelloInterval). The Hello Packet also indicates how often a neighbor must be heard from to remain active (RouterDeadInterval). Both HelloInterval and
RouterDeadInterval must be the same for all routers attached to a common network. The Hello packet also contains the I
P address mask of the attached network (Network Mask). On unnumbered point-to-point networks and on virtual links this field should be set to 0.0.0.0.”

In other word to say, this is how OSPF works as per Hello & Dead interval to form a neighbor-ship with another router… Additionally, to form neighbor-ship with another OSPF router, there is also other fields to be matched in addition of the hello/deal interval which are:
- Area id
- Authentication password
- Stub area Flag

All those are contained in the Hello packet sent from neighbor router(s) they must be matched. I really advise you to check this lesson where you can find more information about this topic:
Introduction of OSPF

Hope I could answer your question.

Hi Rene,

I have two established neighbors, just take dead interval as 40 secs if dr goes down for 30 secs bdr will wait for dead interval to proceed further or will it become as dr ??

Hi Karthik,

If your dead interval is 40 seconds then OSPF will wait for 40 seconds until it declares the neighbor as dead, deleting the neighbor adjacency. If you want to see this in action, connect two routers to each other with a switch in the middle, then disconnect one of the links :slight_smile:

Rene

sir
1> you have this command ip ospf dead-interval minimal hello-multiplier 3 why you have used 3 what is the meaning of minimal hello-multiplier 3 , I understand that when we select minimal the =n our dead interval time will be 1 but why you used hello-multiplier 3 please explain it but if i am using 4 instead of 3 my neighbor adjancecy is working on R1 i have used hello-multiplier 3 and R2 i have used hello-multiplier 4 and they have neighbor adjancency please explain it

1 Like

Hello Bhai,

The “3” in that command means how many hello packets will be sent in 1 second period. Also when you use that command, that is ip ospf dead-interval minimal hello-multiplier x; The x number or the number of hello packets sent per second doesn’t need to match between the two neighbors.

The main important thing about this feature is for critical environment to be able to detect the loss of neighborship in 1 second.

Regards,

1 Like

hi,
do i need to cofigure BFD under router process and under the specific interface?
BDF is very fast to detect link failure, can be generate a problem in a slow network?

Hello Andrea

Take a look at this lesson:


It describes BFD configuration in detail. BFD would not create a problem in a slow network.

Take a look, and if you have any additional questions, let us know!

I hope this has been helpful!

Laz

Hello Rene,
I have a query …
I am two OSPF routers connected via Microwave media as a WAN link … As Microwave links at times are unstable, so we see frequent flaps in OSPF neighborships …
To control or to limit the flapping, I increased the dead-int from default 40 to 300 sec (It’s a linear P2P topology, without any redundant path) … But still moment WAN i/f goes down & come back (<5 sec), OSPF neighborship also goes down and come back …
Why both Routers doesn’t wait for Hold timer to expire before bringing the neighborship down? Although I see other end router still wait for 300 sec before saying nbrship is down …
Moment WAN link goes down, neighborship also see a break at this local router without waiting for Dead-Int to expire …
Is there anyway by which I can bypass WAN link Down status at-least for certain Hold period, so that I can OSPF nbrship not getting down during such frequent flaps on WAN i/f …
Best Regards!
Ashish JAIN

Hello Ajain

Here are a couple of things you should keep in mind as you adjust your config:

During instability, if one router on one end keeps the neighborship up while the other does not, you must take a look and see if there are any differences between the two. Remember that timers should be the same on both ends of the OSPF neighborship.

Also, keep in mind that no matter what timers you configure, if the actual physical interface on that router goes down, OSPF will immediately tear down the neighbor adjacency. Since the interface is down it considers the link lost. So you should check to see if the network instability is indeed due to your WAN link, and is isolated on the microwave equipment, or if it is due to the flapping of the actual physical interface of the router.

There are various other options to use such as OSPF SPF Scheduling and Throttling as well as LSA Throttling, but these are useful in other topologies. Yours is a point to point, and these would not make a difference.

Let us know how your troubleshooting proceeds, and if you have any other questions along the way, let us know so we can help you further.

I hope this has been helpful!

Laz

Hello Laz,
Many Thx for your feedback …
Timers are same on both side & that’s the reason Nbrhsip & adjacency were established during normal conditions …
What I was trying to say before is that moment R1 (MW equipment @ site1) receives i/f Down alarms bec of MW link instability, it immediately tears the nbrship as well. however other end router at times wait for few seconds until it also receives Eth/WAN DOWN alarms or Dead Int Expires which ever is touched first (Dead Int is set to 500 sec in my case) … and thus both routers show Nrbship down until WAN link is restored …
This is very much like i/f shutdown / no- shutdown happening at higher rate because of unstable MW link (WAN link) …
As I don’t have redundant path, so OSPF & LSA throttling won’t be of help… as you also mentioned above …
For a while I am trying to stabilize the WAN link only …
My query was that, is there any way by which we can avoid break in Nbrships when WAN i/f goes down, but now I got that is not possible at the very basic concept …
Initially Hello Int was 10sec & Dead Int was 40 sec but we were observing frequent break in OSPF even though WAN i/f was always UP … Then I noticed, MW link was intermittently generating Error Seconds during poor SNR conditions (radio link fading) … So, to reduce the flaps bec of Error seconds (During ES, there was a possibility of loss of OSPF Hello pkts) , I though to increase the Dead Int from default 40 to 500 sec … and now even though I see Intermittent Error seconds on MW link but OSPF Adjacency still remains up … Only when WAN completely goes down, then I see a break in Nbrship …
My further understanding is that as long as I have a linear topology, increasing the Hold Timer/Dead Int to max supported values will not trigger any side-effects …
Best Regards!
Ashish JAIN

Hello Ajain

Thanks for the details of how you’re getting along with this. I believe that your attempt to stabilize the WAN link is the correct approach to resolving your particular issue. Indeed, since you only have a single path, the changes in the timers will not have any negative effect on the operation of OSPF.

Thanks once again for sharing!

Laz

Hi.

I think there may be a slight mistake.

We can reduce the dead interval to 1 second. If you use the ip OSPF dead-interval minimal command then you also have to specify the hello interval here:

R1(config-if)#ip ospf dead-interval minimal hello-multiplier ?
<3-20> Number of Hellos sent within 1 second

:exclamation: You wrote that when we use the keyword “Minimal” then we have to specify the hello interval, however, in the configuration photo, it’s written “Number of Hellos sent within 1 Second”

:mag: Because I was trying to understand how the “333 msec” come!! and the hello interval doesn’t fit here, the “Number of Hellos sent within 1 Second” does make sense as I believe, correct me if I’m wrong, please.

Thank you for your effort.

1 Like

Hello Ameen

The hello multiplier value does not specify the hello interval directly. It specifies how many hellos are sent in one second. This specifies the hello interval indirectly.

So if you set a hello multiplier of 5, that means 5 hellos per second, which means one hello per 200ms. Therefore the hello interval is 200ms.

Similarly, as in the lesson, the hello multiplier of 3 means 3 hellos per second, which means one hello per 333ms, therefore the hello interval is 333ms.

I understand how this can be confusing since the term “hello interval” was used in the text. I will ask Rene to rephrase it to make it clearer.

Thanks for the feedback!

I hope this has been helpful!

Laz

1 Like

Hi @academicprogress10 ,

I made a change so it’s not confusing anymore. Thanks for sharing your feedback!

Rene

1 Like