Etherchannel trunk configuration questions

[george_daly]

A couple of questions on etherchannel config if I may:

I have an access switch (2960) plugged into the core (6509). VLAN pruning is configured on the 6509 L2 etherchannel but not on the access side, eg I've specified 'switchport trunk allowed vlan 5,6,7,8,9' on the 6509 port-channel but not on the 2960. Would this configuration be considered incorrect/suboptimal? What are the implications of this configuration?

2nd question: These are all gigabit ports but 'show etherchannel <number> detail' output on both sides shows the mode as follows:

Mode = On/FEC

Should this be set to On/GEC for gigabit rather than FastEthernet? If so what are the implications of the mode being set to FEC?

Any replies much appreciated.

George

[amit-singh]

Hi George,

As far as the first question is concern, you should have the same configuration on both sides of the trunk port in an etherchannel configuration. The configuration that is done is not the vlan prunning it is basically removing the vlans manually from the trunk port.

The implication of the above config is that if there is somehow any traffic coming over the vlans which are removed on the Cat6500 side, that traffic will travel all the way to 6500 and get dropped. That means extra traffic will be carried over the trunk ports and hence extra bandwidth untilization on the ports. It has to be set on both the sides of the trunk to save unneccesary bandwitdh untilization.

I really have to look for your second answer. Which module is it. Show version will help

HTH,Please rate if it does.

-amit singh

[george_daly]

Hi Amit,

Thanks very much for your reply.

Indeed I was confusing VLAN pruning with manually clearing the VLANs from the trunk.

I was hoping that may have explained the issue I'm experiencing but it looks like it may not.

The actual problem that I'm having is that the 2960 is dual-homed to two 6509s. When the primary 6509 (STP root bridge) fails the 2960 should fail-over to the secondary 6509(secondary STP root bridge), but that doesn't happen. Instead an etherchannel misconfiguration error occurs on the 6509 and the 2960 is now offline.

If the primary 6509 comes back online the 2960 will not bring the etherchannel to it back up either, even though it was obviously able to successfully establish the etherchannel initially.

If I reboot either the 6509 or the 2960 the etherchannel then comes up Ok.

There are lots of other access / distribution switches which are also multi-homed to both core switches but I only experience this problem with the 2960.

The major differences between the 2960 and the switches that fail over Ok is that the 2960 runs RSTP (as do the 6509s) and has etherchannels to the core whereas the other switches run PVST+ and have unbundled 1Gb uplinks to the core.

Any suggestions on diagnosing the above gratefully received!

BTW I think the 2nd question can be ignored as it was a shot in the dark at resolving the above question but looks like a red herring.