Cisco Support Community
Community Member

Dot1q Trunk Issue

We have 50 2950 edge switches all trunked to a Layer 3 6509 .

The trunks are dot1q trunks all running latest software releases .

Issue relates to not being able to ping the switch, while the trunk is showing up on the 6509 and CDP info is visable .

Bring the trunk down and up again and service is restored .

Any ideas?




Re: Dot1q Trunk Issue

do you allow your managment vlan across the trunks?

Community Member

Re: Dot1q Trunk Issue

Management Vlan is allowed down allow the trunk . The issue was seen on three switches last night , no other instances since.

Community Member

Re: Dot1q Trunk Issue

I have had similar experiences with the same setup. We have 2950's downstream from a 6506 with the MSFC card in it. Every once in a while the switch seems to lose its way to the downstream 2950's. To fix it I have had to get into the MSFC and ping from there to the downstream switch. As soon as I log back into the switch part of the 6500, I am able to ping again. It is like it discovered the router all over again.

Very strange.

Re: Dot1q Trunk Issue

Please check the following:

Is your native vlan set to the same vlan-ID on both ends?

Does this setting match with your default vlan?

With 802.1q, the native vlan is transmitted untagged. Frames may not be recognized correctly when a tagged frame is received, and an untagged frame was expected vice-versa. This can result in losing connections in the way that is described.

It is good practice to set your default vlan as native.

Community Member

Re: Dot1q Trunk Issue

Regarding the "native VLAN transmitted untagged" - I much prefer to have

*all*VLANs tagged, which is possible in recent versions of IOS for 3550

and some other switches ("vlan dot1q tag native"). I believe I also read that

this will cause untagged packets entering a dot1q trunk port to be dropped,

which is a good thing.

CreatePlease to create content