ExtremeSwitching (EXOS)

  • 1.  arp fast detect

    Posted 04-10-2015 08:39
    I have a problem with ARP records on BDx8 and x670 switches
    All ARP records bind to physicall ports.
    If any port goes shutdown appropriate ARP record immediatly removed from db. This behaviour infulent to flow-redirect which has ping health-chek configured and if port flapping flow-redirect immediatly goes down, without waiting timeout interval.

    configure flow-redirect health add nexthop priority 1
    configure flow-redirect health nexthop ping health-check interval 2 miss 6

    show log:
    Flow-Redirect [health] Nexthop [] becomes down by fast detection ARP

    Anybody now how to disable ARP to ports binding and allow to switch to remove ARP record only by configured timeout
    How can i disable "arp fast detection" mechanism?

  • 2.  RE: arp fast detect

    Posted 04-12-2015 14:42
    Does your requirement says.. ARP should not go away even though port went down and it stay there till flow-redirect timeout? If yes, then I don't think we can achieve this.

  • 3.  RE: arp fast detect

    Posted 04-13-2015 06:16
    Hi! Thanks for you commect, Sumit!
    My requirement is that ARP should not go away immediatly after port shutdown, it should remove only by arp timeout like in cisco switches. Flow redirect have only side effect from this behaviour.

  • 4.  RE: arp fast detect

    Posted 05-04-2015 13:14

    You may want to try adding static FDB entries as a workaround or you can mess around with setting the fdb aging time to 0, which would effectively make all FDB entries static (this may cause issues, so be careful).

  • 5.  RE: arp fast detect

    Posted 05-26-2015 08:42
    Thank you guys, i understand that it is internal behaviour. Should think about another possible ways for monitoring ip reachability, may be tcl for example.