3 Replies Latest reply: Apr 15, 2010 5:17 PM by Keith Barker - CCIE RS/Security, CISSP RSS

    Hold Down Timers

    Nabi

      Hi there,

       

      I am studying for the CCNA exam and came across Holdtime timers to stop looping. The definition i got have confused me as to actually how holddown timers works.

       

      " When a router receives an update from a neighbor indicating that a previously accessible network isn’t working and is inaccessible, the holddown timer will start. If a new update arrives from a neighbor with a better metric than the original network entry, the holddown is removed and data is passed. But if an update is received from a neighbor router before the holddown timer expires and it has an equal or lower metric than the previous route, the update is ignored and the holddown timer keeps ticking. This allows more time for the network to stabilize before trying to converge"

       

      Could you please highlight more on this.

       

      My confusion is the holddown timer will start and not accept any updates for that certain period of time default 180secs. When this new update with better metric is recieved the timer is removed and routing tabel updated.

       

      Could you give examples of this better route and how this process actually works.

       

      Cheers

      Nabi

        • 1. Re: Hold Down Timers
          Keith Barker - CCIE RS/Security, CISSP

          Nabi-

           

          When the hold down timer starts, it is as if the router is putting  its hands over its ears, and mumbling "I am not going to believe any new  updates regarding the network that is in hold down".   The reason for  the hold down is to minimize the impact from routers who perhaps didn't  get the message that the network (route) was down and are still  advertising it with a reachable (but not as good as the original source)  hop count.   Sort of like a timeout, until the network has had a better  chance to convergence.   For example, lets say that R3 has been  learning about the 77.0.0.0/8 network from a neighbor (not shown), R3  has been advertising that route to R2 with a hop count (cost) of 7.   When the hold down timer starts for that 77.0.0.0/8 network, R2 will not  believe any new updates from R1 or anyone else (until the hold down  timer expires), unless the new metric is 6 or less (better metric).

           

          Here is a  text diagram, with R1 - R2 - R3 we can use as a reference.

           

          10.0.0.1/8
          e0
          [R1]
          e1
          20.0.0.1/8
          |
          20.0.0.2/8
          e1
          [R2]
          e0
          30.0.0.2/8
          |
          30.0.0.3/8
          e0
          [R3]
          e1
          40.0.0.3/8
          |
          additional  neighbor(s) to R3

           

           

          Even  if the original router who advertised the route comes back on line, and  R3 begins advertising the specific route of 77.0.0.0/8 (which was in  hold down a moment ago) it will still be in hold down, (until the timer  expires), even though R3 is advertising the route with the same metric  (hop count) of 7 that it was originally.   That is the theory behind the  hold down timer.  (The network is considered "possibly down" during  this time, and the router will still forward based on the routing  information it had as it went into hold down.  Packet will still flow  towards what is believed to be the correct direction.)

           

          With RIP, a  neighbor would be a directly connected neighbor.  In this example, R2  has only 2 possible RIP neighbors, R1 and R3.

           

          The hold down timer begins, by default, when an update regarding that route hasn't been seen for 180 seconds.

           

           

          In reality, Cisco modifies code  occasionally, and the actual results will depend on the version of  IOS's reactions to triggered updates, poisoned routes (poison reverse),  etc.   For example, currently, when a poisoned route (16 hops) is advertised by a neighbor, the router today will cancel the hold down timer, and begin sharing the bad news (16 hops) to its neighbors.

           

          Best  wishes,

           

          Keith

          • 2. Re: Hold Down Timers
            Nabi

            Thanks Keith.

             

            I have a better understanding of the concept now.

             

            Cheers

            Nabi

            • 3. Re: Hold Down Timers
              Keith Barker - CCIE RS/Security, CISSP

              Here is a link that goes through an entire debug of the process.   It may be a bit more than you need, but a good reference.

               

              http://blog.ine.com/2010/04/15/how-basic-are-rip-timers-test-your-knowledge-now/

               

               

              Best wishes,

               

              Keith