IEOC - Internetwork Expert's Online Community

I know these are the very basic stuff, but I am still confused about it based on the timeline
I feel dumb ... :-)

RIPv2 Default Timers:

Update: 30 sec
Invalid: 180 sec
Hold Down: 180 sec
Flush: 240 sec

Assume at t=0 sec, a route goes down

  t=0         t=30                    t=180      t=240     
---|----------|--------------------|----------|--------------->

   A            B                         C           D         

At Point A: a route goes down
At Point B: first update timer expired
At Point C: the route is marked as invalid, hold-down timer starts
At Point D: the flush timer kicks in, the route is flushed

So basically, the route will appear in hold-down state for only 60 sec before it is flushed out of the routing table, is that correct?

Thanks!

I had to lab this one up to fully understand it. You can do it in Dynamips with three little routers real easy. 

I cannot recall my exact findings - but here are some quick observations for ya - the update is how often the routing table is broadcast. But remember, RIP will do triggered updates in V2. 

You are on the right track - the timers are cumulative in nature. 

I thought all these timers started from t0. i.e. holdown of less prefered route would enter rib around the same time best route went invalid. Think I need to lab also.

Regards

Von

I read a ton of cisco books, but never seen one that tells you when the RIP timer starts or ends in a timeline fashion. (maybe i was reading too fast)

They will tell you, e.g. the hold-down timer is 180 s or expired in 180 s, and that's it.

Yeah, i should lab it up ...

VirtualGuy....You have it right after 4 mins without and update...poof* the rip route gets deleted.

The confusing thing is the holddown timer...its not really used for anything.

A couple of interesting things at T180

1. The route is marked invalid (that is it is not longer advertised including redistribution) but remains in the local table until the flush time reaches zero.

2. The holddown timer kicks in starting at 180 but only counts down to 120 not zero as you would expect. One way to see this is with the show ip route x.x.x.x command between T180 and T240.

Well, The Holddown timer does not start at T180 but it expires at T180.

Every update packet received resets the holddown timer after every 30 seconds. If the update is not received for 180 seconds, the route becomes invalid and is not included in the updates sent by the local router. The local routing table starts showing it as possibly down for another 60 seconds. That is from T180 to T240.

And as T240 is reached, the routers flushed from the routing table.

Like I stated before...hold down is the confusing one!

Hold down timer does not start until a route with a worse metric is recieved.  If a route is lost and no additional updates are recieved hold down does not start until T180 when the local router poisons the route (that is route update happens).

hold-down in Cisco's RIP is the idea of not accepting a route update until a certain about of time has elapsed.  When an advertisement is received with a worse metric wait or hold for 180 seconds before installing in the routing table.  (it a similar to the Feasibility condition is EIGRP a worse route could mean a routing loop)

Hold down is not something that gets constantly reset because it only comes into play when triggered and the trigger is receiving a route advert with a higher metric.

Now the normal rip timers (update, invalid and flush) are all constantly running and updated but hold time is a special case that runs upsome independently.

Don't you love it when Cisco enhances a protocol!!!