spanning-tree / Err-Disable finding the loop via software

Discussion in 'Cisco' started by colin, Mar 30, 2010.

  1. colin

    colin Guest

    Hi NG,

    i need some advice/comments on finding an assumpted L2 Loop via software...
    instead of tracking it with shutting down the redundancy the hard way..

    Setup of szenario:

    3550 gi0/2 -- gi2/2 4506 -- PrtChan -- PrtChan -- 6509


    Cat 4506 and Cat 6509 play HSRP whereby the 6509 is always the prio 105 of
    all configured vlans.

    On the Cat 3550, witch is only attached to the 4506 via gi2/2, i get
    following messages:

    Mar 30 14:16:10: %ETHCNTR-3-LOOP_BACK_DETECTED: Keepalive packet loop-back
    detected on GigabitEthernet0/2.
    Mar 30 14:16:10: %PM-4-ERR_DISABLE: loopback error detected on Gi0/2,
    putting Gi0/2 in err-disable state
    Mar 30 14:16:11: %LINEPROTO-5-UPDOWN: Line protocol on Interface
    GigabitEthernet0/2, changed state to down
    Mar 30 14:16:12: %LINK-3-UPDOWN: Interface GigabitEthernet0/2, changed state
    to down

    Thats the 3550 gone to Err-Dis!

    on the 4506 i get following logs:
    23 08:39:53: %C4K_EBM-4-HOSTFLAPPING: Host 18:A9:05:62:4E:DA in vlan 100 is
    flapping between port Gi2/6 and port Po10
    Mar 23 08:39:53: %STANDBY-3-DUPADDR: Duplicate address 192.168.6.3 on
    Vlan60, sourced by 000e.383e.483f
    Mar 23 08:39:54: %C4K_EBM-4-HOSTFLAPPING: Host 18:A9:05:62:4E:DA in vlan 100
    is flapping between port Gi2/6 and port Po10
    Mar 23 08:39:54: %C4K_EBM-4-HOSTFLAPPING: Host 00:D0:03:EA:E0:0A in vlan 110
    is flapping between port Po10 and port Gi2/6
    Mar 23 08:39:54: %C4K_EBM-4-HOSTFLAPPING: Host 18:A9:05:62:4E:DA in vlan 100
    is flapping between port Gi2/6 and port Po10
    Mar 23 08:39:54: %C4K_EBM-4-HOSTFLAPPING: Host 00:D0:03:EA:E0:0A in vlan 99
    is flapping between port Gi2/6 and port Po10
    Mar 23 08:39:54: %C4K_EBM-4-HOSTFLAPPING: Host 00:00:00:00:FE:01 in vlan 100
    is flapping between port Po10 and port Gi2/6
    Mar 23 08:39:54: %C4K_EBM-4-HOSTFLAPPING: Host 00:D0:03:EA:E0:0A in vlan 200
    is flapping between port Po10 and port Gi2/6
    Mar 23 08:39:54: %C4K_EBM-4-HOSTFLAPPING: Host 00:1E:0B:C3:23:78 in vlan 100
    is flapping between port Po10 and port Gi2/6
    Mar 23 08:39:54: %C4K_EBM-4-HOSTFLAPPING: Host 00:0E:7F:FD:58:B0 in vlan 100
    is flapping between port Po10 and port Gi2/6
    Mar 23 08:39:55: %C4K_EBM-4-HOSTFLAPPING: Host 00:D0:03:EA:E0:0A in vlan 99
    is flapping between port Po10 and port Gi2/6
    Mar 23 08:39:55: %C4K_EBM-4-HOSTFLAPPING: Host 00:90:FB:24:BA:00 in vlan 100
    is flapping between port Gi2/6 and port Po10
    Mar 23 08:39:55: %C4K_EBM-4-HOSTFLAPPING: Host 00:1F:29:69:9D:6E in vlan 100
    is flapping between port Gi2/6 and port Po10
    Mar 23 08:39:55: %C4K_EBM-4-HOSTFLAPPING: Host 00:D0:03:EA:E0:0A in vlan 100
    is flapping between port Gi2/6 and port Po10
    Mar 23 08:55:58: %C4K_EBM-4-HOSTFLAPPING: Host 00:50:56:83:45:A3 in vlan 100
    is flapping between port Po10 and port Gi6/5
    Mar 23 08:56:59: %C4K_EBM-4-HOSTFLAPPING: Host 00:50:56:83:45:A3 in vlan 100
    is flapping between port Po10 and port Gi6/5

    the interessting one i guess is this:
    Mar 23 08:39:53: %STANDBY-3-DUPADDR: Duplicate address 192.168.6.3 on
    Vlan60, sourced by 000e.383e.483f

    when i issue the following on the 4506

    asw03#sh mac-address-table | i 000e.383e.483f
    12 000e.383e.483f static ip,ipx,assigned,other Switch
    30 000e.383e.483f static ip,ipx,assigned,other Switch
    50 000e.383e.483f static ip,ipx,assigned,other Switch

    it sees its own mac addr, what a surprise!
    so i do guess i got some bad things happening on L2, loopin all over.


    ------------------------------------------------------------
    as i just tried following within a complete seperate lab:

    3560 fa0/2 -- fa0/1 3550 from 3550 to itself from fa0/47 to fa0/48
    (looped)

    on the 3550 i configured: no spanning-tree vlan 1-999 making sure
    it runs as if i would use some unmanaged switch, thereby creating a loop on
    it self.

    within this LAB it was a pretty ugly szenario, since i actually thought the
    3560 on fa0/2 would track it and STP would go into the Blocking state on
    fa0/2, but it did NOT.
    so the hole thing looped and flashed like an x-mas tree on fa0/2 aswell of
    course on the 3550.

    so i am considering in order to find the loop to go and take the approach
    via storm-control, thereby issuing the action trap mode, in order not to
    shutdown my hole network.

    on my Distribution Trunks/Uplinks towards the Access-Switches:

    storm-control broadcast level pps 500
    storm-control multicast level pps 500
    storm-control action trap

    i guess this should give me an idea from witch link i get alot of
    broadcasts..

    i still can't explain why my 3550 in the real world setup went down with
    Err-Dis, since it is single attached to the 4506....

    so if you got other idea's how to track L2 Loops without pulling out the
    redundancy until it stops, feel free to contact me and send me youre advice.

    Thank you
    colin
     
    colin, Mar 30, 2010
    #1
    1. Advertisements

  2. colin

    Thrill5 Guest

    This problem has nothing to do with spanning tree. A very long time ago
    (probably around 2001 or 2002) I had a similar problem on 3550's and after
    much troubleshooting I found it was software bug on the 3550's. The
    workaround was issuing a "no keepalive" on each of the uplink interfaces on
    the 3550's. After a quick search, I found a reference to this bug id
    CSCea46385.
     
    Thrill5, Apr 1, 2010
    #2
    1. Advertisements

  3. colin

    colin Guest

    Hey Thrill,

    you just saved me alot of time!!

    BIG THANK YOU!

    Regards
    Colin
     
    colin, Apr 1, 2010
    #3
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.