Hi All,
I am studying for CCIE SP Lab, and also use TE tunnels in the network I work on for the day job. I watched the CCIE SP videos Brian did, and noticed that when he had a TE tunnel with an explicit path as preferred and a dynamic as second choice the tunnel failed between the paths within 1 second (1 ping dropped) when he broke a link the explicit path was using. However when I tested this in a lab scenario the failover takes 5 seconds. I'm trying to understand what takes that 5 seconds. So the test is essentially that I fail a link that the explicit path uses and monitor the ping drops from the CE-CE perspective. On the link that fails I have configured BFD on the ISIS process, this hasn't caused any noticeable change in how quickly the TE tunnel reroutes, although that makes sense bearing in mind TE uses RSVP. What exactly is taking RSVP 5 seconds? Presumably to report back to the tunnel head end that a link has failed, but I don't understand why in Brians video this was 1 second? Are there timers you can tweak on links?
N.B i understand FRR is a way around this, but in a live network I may not wish to use FRR to reduce complexity. If the TE tunnel can failover between paths in 1-2 seconds this may be acceptable to most customers, but 5 seconds is too slow.
Thanks for any help,
Sam