port-channel issue

Unanswered Question

I have a Cisco 6509, where we are doing port channels/aggregating ports to a HP chassis. In any case during the initial setup we made some extra port groups. And being the good admin I wanted to clean up the extra port channels.

Well right after I took the ports out of the port-channel all hell broke loose. And the switch stopped traffic everywhere.

I think it was around the time when I put a command no channel-group while in a range of ports. By the way these ports only go to a HP chassis nothing was running on it at the time. I am attaching the log from the putty. The only thing after the fact that I noticed is that I did a no channel-group without a channel-group number. But I would hope that it wouldn't crash from that and at worst just not accept the command. Any thoughts are more then welcome. But the real question is whether what I did caused it or did something else happen. I am running IOS 12.2 and other then another port channel to another HP chassis we don't have any others. And it took out all network traffic not just to the port channels.

I have this problem too.
0 votes
  • 1
  • 2
  • 3
  • 4
  • 5
Overall Rating: 3 (1 ratings)
nnandrek Mon, 06/16/2008 - 20:18


Are those L2 port channels i.e. port channels on a switchport? Are all the portchannels in the same broadcast domain?

If so, when you removed the portchannel from the interface range, did you wait for spanning tree to re-converge? atleast 30 seconds to see if traffic resumed? If those port channels were in the same broadcast domain, removing them will cause spanning tree topology change and all ports in the same broadcast domain (same vlans) will stop forwarding for some time. will all move to listening state. Please wait for around a minute and see if traffic resumes. See the spanning tree details for the vlan to see if the state of the ports.



(please rate if the posts help)

Thoae are port channels on switchports. As we were just aggregating the ports together:

Here are the commands we used to setup the port groups. I am just including this to clarify things, as I'm not an expert at this. As you can see we aren't setting up spanning trees as far as I can tell.

int range g3/37-40

switchport mode trunk

switchport trunk allowed vlan 1-500

channel-group 11 mode active

int po11

no shut

Edison Ortiz Tue, 06/17/2008 - 05:20

Spanning-Tree is enabled by default on Cisco Switches, you can verify with the command show spanning-tree. What the previous poster was implying, if you have multiple switches in your network, one switch will be the Spanning-Tree Root for a Vlan or multiple Vlans.

Any changes on the switchport connectivity between the switches can cause a Spanning-Tree recalculation. This process is similar to a routing process (OSPF or EIGRP) after losing a neighbor or link.

During this time, all operations (packet forwarding) in the switch are halted and that's probably what you saw as a 'crash'.

Always make sure to schedule changes during a maintenance window (off-hours) to avoid a network outage such as the one you experienced. I don't think you needed to reboot the switch. Waiting for 60 seconds or so should be sufficient.




Well after looking around we do have spanning tree enabled. Also the network wasn't responsive for about 5-10 minutes -which is about when I rebooted it. We do have about 40 switches or so, so it could have taken awhile between all of them.

I should have done a show tech support as I would have had more information. Other then a show tech support should I do anything else to get information before trying to resolve a problem?

Edison Ortiz Tue, 06/17/2008 - 05:54

5-10 minutes is quite a long time.

You needed to identify what wasn't responsive.

- Was it a partial, total outage?

- Were you able to ping from devices within the same subnet ?

- Were you able to ping to devices in another subnet ?

- What was the status of the switchports ?

- Is the switch serving as a Layer3 devices for the network, if not - was the connection to the Layer3 device disabled ?

Without knowing your network topology, is very hard to instruct you on how to troubleshoot a network outage. Troubleshooting comes from years of experience in the field. Configuring a network is much easier since you can follow a documentation guideline.

As for the show tech-support query, yes - the command is very useful if you are planning to open a TAC case or if you want to have a complete hardware state of the device when the event occurred.





This Discussion