2

We have two links for a site , both the links are running OSPF , if one of the link keeps flapping the network at site becomes unstable. How do it make the flapping link down automatically

4 Answers 4

1

If the link flaps OSPF will run the SPF algorithm afresh to re-calculate the shortest path to all available destinations.

you can change the throttle numbers to make the network more stable

The default timers are below

  • Initial SPF schedule delay 5000 msecs

  • Minimum hold time between two consecutive SPFs 10000 msecs

  • Maximum wait time between two consecutive SPFs 10000 msecs

SPF calculations occur at the interval set by the timers throttle spf command. The wait interval indicates the amount of time to wait until the next SPF calculation occurs. just tune those timers to force OSPF recalculate

Have a look at this link on how to change it. OSPF throttle tune
and for very useful TS command OSPF TS

1

Tuning OSPF timers is a good idea, but there are methods for shutting down a flapping interface on lower levels. For example, Cisco switches have a feature which puts a port into err-disabled state thus effectively stopping transmission over this interface is it flaps more that 5 times in 10 seconds (http://www.cisco.com/c/en/us/support/docs/lan-switching/spanning-tree-protocol/69980-errdisable-recovery.html)

Moreover you can you can tune the defaults of this feature using info from this post https://supportforums.cisco.com/document/25101/how-configure-errdisable-flap-setting-values-cisco-catalyst-switches-running-cisco

4
  • putting interface in err-disable mode need you perform shut and no shutdown each time this port enter the err-disable mode . which is non-automatic process .sure you don't need to restart the interface each time flapping occurred.dynamic routing protocols used to handle topology change automatically
    – Gadeliow
    Commented Sep 14, 2015 at 22:48
  • There are mechanisms for automatic reenablement of errdisabled ports like: #errdisable recovery cause link-flap ** and **#errdisable recovery interval 400 for setting autorecovery interval to 400 seconds. Commented Sep 15, 2015 at 7:30
  • other big problem i duet if you are a ware of or not . Errdisable is a feature that automatically disables a port on a Cisco Catalyst switch not Cisco Router and mainly you run OSPF on routers . here you can find a list of supported plateforms cisco.com/c/en/us/support/docs/lan-switching/…
    – Gadeliow
    Commented Sep 15, 2015 at 12:56
  • By flaps, do you mean subsecond flaps? Check out the link-debounce timer concept. You can use that to delay the reporting of such flaps if it happens within a set interval. But your question is about bringing it down permanently - and for that I agree with @Gadeliow
    – ajaysdesk
    Commented Sep 30, 2015 at 2:57
1

i would think that something like "ip event dampening" might be a better fit. when tuned correctly, it should hold a flapping interface down and when it's stable again allow it to come up.

this has long been available on Cisco IOS and more recently on JunOS.

1
  • ÏP Event Dampening only identifies down or up in a physical interface, not in a logical level, like an interferance causing the changes of routing tables
    – user24949
    Commented Apr 27, 2016 at 19:57
0

You may use IP Event Dampening to solve your issue. For each flap a so-called penalty is incremented and when the penalty reaches the suppress threshold the interface is being dampened to ensure network stability.

IP Event Dampening is an automatic process that uses timers to make sure an interface is stable before enabling it again.

To get more information on IP Event Dumpening click here

Not the answer you're looking for? Browse other questions tagged or ask your own question.