vlan translation - cheat

Hi,

This is my setup: Site a: interface GigabitEthernet5/2 switchport switchport trunk encapsulation dot1q switchport trunk native vlan 2500 switchport trunk allowed vlan 444,333 switchport mode trunk

site b: interface Gi0/1 switchport trunk encapsulation dot1q switchport trunk native vlan 123 switchport trunk allowed vlan 444,333 switchport mode trunk interface Fa0/1 switchport access vlan 123 switchport mode access

what im trying to do, is have a edge port in vlan 123, go out a trunk port (set as native vlan so it should strip its tag) come in on the other port untagged (in native vlan 2500) so it should be tagged as

2500. hence it should be a fairly straight forward one to one vlan translation.

site b switch doesnt support system extended vlan ID's and that particular vlan is a must. anyways, on switch A I have 5 mac addresses in that vlan, on switch B, doing show mac-address-table int g0/1 I can't see any mac's from anything other than 444,333 so it doesnt appear to be working.

any ideas?

Flamer.

Reply to
die.spam
Loading thread data ...

I think one of the issues is that im running PV-RSTP. so if I dont "trunk allow" that vlan on the port that vlan is forwarded on that port. and if I do add it on the port at the same time as its a native vlan im not sure it will still add/strip the dot1q tag.

Reply to
die.spam

Well I have some ideas but I don't know if they might help.

I don't know exactly how PV-RSTP works.

Your plan seems reasonable, well a terrible hack but I have done worse I am sure:-)

If you have a look at sh int switchport

and for each of the vlans sh span vlan xxx

You might see something interesting - or not.

Might you need switchport nonegotiate !? or is it sw mode trunk noneg !?

If you don't disable CDP on the interfaces you will forever get native VLAN missmatch messages filling the logs.

At one time some switches did VLAN translation. I never used it so I can't recall the details.

If there is a reason it won't work the following will as long as VLAN123 is available for use on the site a: side too. i.e. not used for something else.

Just trunk 123 across and then create an access port on 123 and another on 2500. Connect with cable. That is cable connected to two ports on same switch. "Wastes" two ports but maybe that does not matter.

Obviously you may need a crossover cable or auto mdix or something. Once again turn off CDP. Oh and mention your disgraceful hack to someone before it causes too much headscratching.

Reply to
bod43

I'm not positive, but you might need to add the native VLAN to the allowed VLAN list.

Reply to
Thrill5

ok I found the native vlan mistmatch error reporting on the switch is coming from STP not CDP, I disabled CDP and still get the message:

%SPANTREE-SP-2-RECV_PVID_ERR: Received BPDU with inconsistent peer vlan id 123 on GigabitEthernet5/2 VLAN2500.

I also added the native vlan's to the allowed vlan list which seems to be the only way to make it work. I think two access ports with a x- over may be the next option. The switch B is a 6509 and can support vlan mapping but have not been able to get that to work.

Flamer.

Reply to
die.spam

I would also guess that you need the native VLAN in the allowed list. I did not notice that. I have always just put it in.

Bet that is why it is not working. :)

You might get the same problem STP with the crossover. If it is only a logging issue you might decide to live with it.

As I said I do not know how the new fancier STPs work.

If you don't need STP you could consider turning it off. If you do need it (i.e. have loops) you can't of course.

Reply to
bod43

..... straight forward one to one vlan translation....

On the 6500s there is a "switchport vlan mapping" command which does just that, and might be usable for your situation.

Unfortunately support is dependent on the module port asic and it seems it is not usable on most of the newer hardware. If your hardware does support it : it works per port asic, meaning the mapping will be applied to ALL ports on that asic. Just how many ports that would be is again hardware dependent, but there is an "enable" command per port. I've used this once and it worked just fine, and yes, I've also used external links between two ports.

Luck, Frans

Reply to
brink

Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.