Etherchannel config problem or HW problem?

[Luis Ruiz Lopez]

Hi all,

Hope you can help me out to figure out whether I have something wrong with the config of this etherchannel or if it's a hardware problem. Everything looks like a HW problem to me but maybe you can show me a different direction to handle this issue.

I have 2 Cisco Catalyst WS-C4006 running IOS version cat4000-i5s-mz.122-25.EWA14.bin with the following modules installed:

Mod Ports Card Type                              Model              Serial No.

---+-----+--------------------------------------+------------------+-----------

1     2  Supervisor III 1000BaseX (GBIC)        WS-X4014          

2     6  1000BaseX (GBIC)                       WS-X4306-GB       

3     6  1000BaseX (GBIC)                       WS-X4306-GB       

4    24  10/100BaseTX (FX-MT)                   WS-X4124-FX-MT    

5    24  10/100/1000BaseT (RJ45)                WS-X4424-GB-RJ45  

I have configured ectherchannel 6 as the trunk link between both switches (same config in both switches) with the following config:

interface Port-channel1

switchport

switchport trunk encapsulation dot1q

switchport mode trunk

storm-control broadcast level 5.00

storm-control action trap

spanning-tree guard root

and interfaces gig1/1 and gig1/2 with the following config (same config in both switches):

interface GigabitEthernet1/1

switchport trunk encapsulation dot1q

switchport mode trunk

udld port

no snmp trap link-status

storm-control broadcast level 5.00

storm-control action trap

channel-group 6 mode on

interface GigabitEthernet1/2

switchport trunk encapsulation dot1q

switchport mode trunk

udld port

no snmp trap link-status

storm-control broadcast level 5.00

storm-control action trap

channel-group 6 mode on

I have changed the etherchannel mode to desirable and the same and I aslo tried with LACP instead of Pagp.

The story is that when I enable etherchannel 6, all vlans start flapping and I have to either shutdown inter gig1/2 to be able to leave the net stable or let UDLD do its job and remove the gig1/2 out of Po6 by placing the port in err-disable state mode when the loop is detected.

Here is the messages that I got:

Aug 31 03:34:03  %C4K_EBM-4-HOSTFLAPPING: Host 9C:AF:CA:94:17:90 in vlan 333 is flapping between port Po6 and port Gi2/5

Aug 31 03:34:06  %C4K_EBM-4-HOSTFLAPPING: Host 00:26:73:18:B4:7C in vlan 6 is flapping between port Po6 and port Gi2/4

Aug 31 03:34:06  %C4K_EBM-4-HOSTFLAPPING: Host BC:30:5B:B4:F5:09 in vlan 5 is flapping between port Po6 and port Gi2/6

Aug 31 03:34:11  %C4K_EBM-4-HOSTFLAPPING: Host 00:0A:B7:7D:49:FF in vlan 3 is flapping between port Po6 and port Gi2/5

Aug 31 03:34:12  %C4K_EBM-4-HOSTFLAPPING: Host 00:1E:4F:E3:14:C4 in vlan 6 is flapping between port Po6 and port Gi2/3

Aug 31 03:34:12  %C4K_EBM-4-HOSTFLAPPING: Host 00:0A:B7:7D:49:FF in vlan 6 is flapping between port Po6 and port Gi2/2

Aug 31 03:34:13  %C4K_EBM-4-HOSTFLAPPING: Host 00:1A:A0:55:60:1E in vlan 5 is flapping between port Po6 and port Gi2/6

Aug 31 03:34:18  %C4K_EBM-4-HOSTFLAPPING: Host 9C:AF:CA:94:17:90 in vlan 333 is flapping between port Gi2/5 and port Po6

Aug 31 03:34:23  %C4K_EBM-4-HOSTFLAPPING: Host 00:0A:B7:7D:49:FF in vlan 3 is flapping between port Po6 and port Gi2/5

Aug 31 03:34:25  %C4K_EBM-4-HOSTFLAPPING: Host 00:0A:B7:7D:49:FF in vlan 5 is flapping between port Po6 and port Gi2/2

Aug 31 03:34:25  %C4K_EBM-4-HOSTFLAPPING: Host 00:0A:B7:7D:49:FF in vlan 6 is flapping between port Po6 and port Gi2/2

Aug 31 03:34:32  %C4K_EBM-4-HOSTFLAPPING: Host 00:0A:B7:7D:49:FF in vlan 333 is flapping between port Po6 and port Gi2/5

Aug 31 03:34:33  %C4K_EBM-4-HOSTFLAPPING: Host 9C:AF:CA:94:17:90 in vlan 333 is flapping between port Po6 and port Gi2/5

Aug 31 03:34:37  %C4K_EBM-4-HOSTFLAPPING: Host 00:1E:4F:E3:17:12 in vlan 6 is flapping between port Po6 and port Gi2/3

I tried configuring interfaces gig 5/20 and 5/21 into Po6 and the problem doesnt occur. It is only happening with inter gig 1/2.

If I put gig1/1 and gig1/2 to work together then the flapping starts.

If I put gig1/1 to work alone the flapping stops

If I put gig1/2 to work alone the flapping starts.

And finally two different gig interfaces working together work fine.

So the problem seems to me like a HW problem. Any suggestions?

P.S. I'm using fiber links and one of the supervisor cards has already been changed and also different fibers have been used to discard problem with the fiber.

Thanks a lot for your support and ideas and sorry for this long message!!!

[IAN WHITMORE]

If you're saying thay UDLD kicks in then you have a unidirectional link on a fibre...which means either one of the pairs of one fibre isn't working, or the fibre SFP is not working right.

So your best bet is to switch the SFP's and see if that works (as you said you already checked and switches the fibre itself).

HTH,

Ian

[Luis Ruiz Lopez]

I actually changed the fibers 2 times so I guess this means the problem should not be on the fibers. Apparently the supervisor card has been changed in both switches but that is something that I need to make sure of.

The fiber that was used to connect gig5/20 and 5/21 and that worked well is the one in use now for gig1/1 and 1/2.

Do you think thats the only thing that could be affecting the well functioning of the etherchannel?

Thanks a lot for your very quick reply!!

Luis

[IAN WHITMORE · Accepted solution]

Well yes and no. I just realised. You specify channel-group 6 mode on, but your configuration is for port-channel 1.

You should have a port-channel 6.

If that doesn't fix it, it will be a dodgy SFP (mini-gbic) by the sounds of it.

HTH,

Ian

[Luis Ruiz Lopez]

oh well, sorry about that, it was a mistake.. Thats is because at first I was using Portchannel1 and of the things I did was to create portchannel 6 with the same configuration as Portchannel1 and tried but I'm using port channel 6.

I'll try few things with the SFPs to see if I can fix it and get back to you.

Thx

[IAN WHITMORE]

OK matey. Hope it works out for you.

Regards,

Ian

[Luis Ruiz Lopez]

Hi Ian,

Unfortunately the problem persist.I have changed the GBICs, fibers and still the same.

When I check UDLD status on both switches, the problem is always in ST02 interface gig1/2 even when I swap the GBICs from ST02 to SR03.

If its not the GBIC, what else could be?

Any other idea or sugestion?

Thanks a lot

Luis

[Leo Laohoo]

Move the interface to a different port.  If the issue goes away then you've got a BIGGER issue.

[Luis Ruiz Lopez]

I've tried configuring the etherchannel using interfaces gig 5/20 and 5/21 (SFPs) and the etherchannel works ok. It only happens with int gig1/2 (GBIC).

So you also think it is more a HW problem? Apparently the supervisor card III was also replaced on that switch some time ago.

What is your idea?

[Leo Laohoo]

So you also think it is more a HW problem? Apparently the supervisor card III was also replaced on that switch some time ago.

Indeed.  HW issue.

Can you reboot the chassis?  Sometimes if there's an issue with the port it will show up during the POST.

[Julio Garcia]

can you do #sh interface gi  x/x  capabilities

and compare the two outputs  , check if 'channel' section has 'yes'

[Luis Ruiz Lopez]

I' ve issued the command on both interfaces and the same ouput is given:

GigabitEthernet1/2

  Model:                 WS-X4014-Gbic

  Type:                  1000BaseSX

  Speed:                 1000

  Duplex:                full

  Trunk encap. type:     802.1Q,ISL

  Trunk mode:            on,off,desirable,nonegotiate

  Channel:               yes

  Broadcast suppression: percentage(0-100), hw

  Flowcontrol:           rx-(off,on,desired),tx-(off,on,desired)

  VLAN Membership:       static, dynamic

  Fast Start:            yes

  Queuing:               rx-(N/A), tx-(1p3q1t, Sharing/Shaping)

  CoS rewrite:           yes

  ToS rewrite:           yes

  Inline power:          no

  SPAN:                  source/destination

  UDLD:                  yes

  Link Debounce:         no

  Link Debounce Time:    no

  Port Security:         yes

  Dot1x:                 yes

  Maximum MTU:           9198 bytes (Jumbo Frames)

  Multiple Media Types:  no

  Diagnostic Monitoring: no

When I reloaded the routers one of the times and I think one of the interfaces was blinking with an orange light.