EIGRP Hold Timer

Alceo

I've a bit of confusion,
let's say that R1 has an hold timer of 10 secs and R2 has and hold timer of 20 secs.
Does this means that R2 will wait 10 secs for a Hello from R1 before considering him dead?
And R1 will wait 20 secs for a Hello from R2?

joetest

Yes. R1: I'll send you my Hold-time and if you don't hear from me again before it expires you should consider me as down.

james43026

Alceo wrote: »

I've a bit of confusion,
let's say that R1 has an hold timer of 10 secs and R2 has and hold timer of 20 secs.
Does this means that R2 will wait 10 secs for a Hello from R1 before considering him dead?
And R1 will wait 20 secs for a Hello from R2?

Not quite. If R1 has a hold timer of 10 secs, then it will wait 10 seconds for a hello packet, if none is received it will consider the adjacency between itself and R2 to be down, R2 would wait 20 seconds for a hello packet before doing the same. Default hold time on EIGRP is 15 seconds, with a hello interval of 5 seconds.

koz24

james43026 wrote: »

Not quite. If R1 has a hold timer of 10 secs, then it will wait 10 seconds for a hello packet, if none is received it will consider the adjacency between itself and R2 to be down, R2 would wait 20 seconds for a hello packet before doing the same. Default hold time on EIGRP is 15 seconds, with a hello interval of 5 seconds.

I don't think this is true. In EIGRP, the hello instructs your router, whereas the hold instructs the neighbors. So, if R1 has a hold of 10, then R2 will wait 10 seconds before considering the neighbor to be down. Likewise, if R2 has a hold of 20, then R1 will wait 20 seconds before it considers the neighbor down. On an ethernet segment if you had say R1,R2,and R3 and R1 set it's hold time to say 15, then both R2 and R3 would wait 15 before considering R1 down. OP is correct.

james43026

Not sure how so many people could be confused about this. How would an internal timer affect the settings on a neighbor? The following article may help everyone understand how a hold timer works

Hold-down timer:
This is the amount of time that a router will consider a neighbor alive without receiving hello packets. The hold time is typically three times the hello interval. You can adjust the hold time with the ip hold-time eigrp command.
Unlike OSPF changing the hello timer does not automatically adjust the hold down timer.
Sometimes hello packets are lost in congested networks and neighbor relationships start to flap; in such cases you may want to increase the hold-down time.
Increasing the hold-down timer delays convergence, this is an undesirable effect. If the default hello/hold time value is not suitable for your network try to find the value that best suits your network.

A hold timer specifically tells the router how long to wait without hearing a hello from a neighbor, before considering a neighbor down and no longer available. How would configuring a hold timer on one router, change the hold timer on another router? If you can explain this with proof that would be greatly appreciated.

james43026

It appears I misunderstood the question. I'm sorry, the hold timer configured on an interface is advertised to neighbors on that interface and the neighbors will use this to determine when the router is down.

Alceo

Now I'm reading Lammle's book and I found this in the EIGRP chapter:
"Interestingly, the hold timer configuration doesn’t determine how long a router
waits before it declares neighbors dead; it establishes how long the router will tell others to
wait before they can declare it dead. This means that the hold timers on neighboring routers
don’t need to match because they only tell the others how long to wait"

joetest

james43026 wrote: »

It appears I misunderstood the question. I'm sorry, the hold timer configured on an interface is advertised to neighbors on that interface and the neighbors will use this to determine when the router is down.

no reason to make it so complicated - this is the correct answer.