In this lab I use a Dell Force10 S25P switch [FTOS 8.4.2.6] and a Cisco C3750 switch [IOS 12.2(25)SEE3] to demonstrate 802.1Q trunking.
VLANs are local to each switch’s database, and VLAN information is not passed between switches. If it is desired for nodes on different VLANs on a switch to communicate to other nodes on the same VLANs on other switches via a single or limited amount of links, trunking will need to be configured. Trunk links are used to carry frames that belong to multiple VLANs between nodes over the same link.
A node can determine which VLAN the traffic belongs to by its VLAN identifier. The VLAN identifier is a tag that is encapsulated with the data. Cisco uses two types of encapsulation technologies – ISL (Inter-Switch Link) and IEEE 802.1Q. If you are using switches from vendors other than Cisco, you will need to use 802.1Q encapsulation for trunking; this is an open standard and even Cisco has moved more towards 802.1Q encapsulation.
The way that trunking is configured and the default behavior is totally different on Cisco and Dell Force10 switches; I have seen this cause a lot of confusion for customers using both switches in a heterogeneous environment. With Cisco, when you create a trunk, all VLANs are automatically tagged and transmitted down the trunk link. However, you can remove a VLAN from a trunk with the “switchport trunk allowed vlan remove [VLAN ID]” command. Conversely, on Dell Force 10 switches, a trunk link only sends VLAN traffic that has been explicitly tagged to be sent down the respective interface (ex: tagged gigabitethernet 0/1).
A second point of confusion is the native VLAN. The term “native VLAN” is really only valid when trunking is involved. The native VLAN is basically the VLAN traffic that will be sent down the trunk link automatically as untagged. In Cisco’s case, the native VLAN by default is VLAN 1 so all VLAN 1 traffic will be sent down the trunk link as untagged. Again, you can also change the native VLAN via the “switchport trunk native vlan [VLAN ID]” command.
On Cisco switches the default VLAN is VLAN 1 which also happens to be the default native VLAN. On Force10 switches, the default VLAN is VLAN 1 but the native VLAN is configured via a mechanism called a “hybrid port”. On a hybrid port, the native VLAN is the VLAN of which the port is an untagged member.
So let’s take VOIP (Voice over Internet Protocol) as an example. On a Cisco switch the VOIP phone may be added to say VLAN 100, and the native VLAN of the trunk could be set to VLAN 10. Then the Cisco switch would connect to the VOIP phone which would then connect to the PC as follows: Cisco Switch <----> VOIP phone <----> PC.
In the same scenario, the Force10 switch would be setup as below with the port on the switch connecting to the VOIP phone configured as a hybrid port.
VOIP Setup Using a Dell Force10 Switch
Now that we understand the basics, let’s look at the below lab setup displaying the Dell Force10 S25P switch connected to the Cisco C3750 switch via Port Channel 1.
Lab Diagram
I have changed the default VLAN on the Force10 S25P via the following config mode commands:
– default-vlan disable
– default vlan-id 5
Next, on the Force10 S25P, I configure Port Channel 1 as a hybrid port in interface configuration mode and add the respective interfaces via the below commands:
– portmode hybrid
– channel-member GigabitEthernet 0/21-22
– sw
– no shut
– exit
– interface gigabitethernet 0/21
– no shut
– interface gigabitethernet 0/21
– no shut
Finally, on the Force10 S25P, I untag VLAN 1 on PO 1 and tag it on VLAN 2 and give VLAN 1 an IP address of “10.10.10.1/24”.
(in configuration mode)
– int vlan 1
– untagged Port-channel 1
– ip address 10.10.10.1/24
– no shut
– int vlan 2
– tagged Port-channel 1
– no shut
– end
– write
The following shows the output of the “show vlan” command on the Force10 S25P switch:
Force10 S25P - “show vlan” output
Now I need to configure the Cisco side. Here I also use the same Port Channel, PO 1, however, in this case I use interfaces gigabitEthernet 1/0/1 and gigabitEthernet 1/0/2. First, I create the static Port Channel.
(in configuration mode)
– interface gigabitEthernet 1/0/1
– channel-group 1 mode on
– no shut
– exit
– interface gigabitEthernet 1/0/2
– channel-group 1 mode on
– no shut
– exit
Second, on the Cisco C3750, I create the trunk:
– interface port-channel 1
– switchport trunk encapsulation dot1q
– switchport mode trunk
– no shut
Finally, on the Cisco C3750, I put an IP address on VLAN 1. Interestingly, on Force10 switches, you cannot configure an IP address on the default VLAN which also happens to be VLAN 1. In our case, we changed the default VLAN so we were able to do so. Also, note that VLAN 1 is the native VLAN so, if I ping the Force10 S25P switch, the packets should get sent across the trunk link (PO 1) as untagged packets and received on the Force10 switch as untagged packets.
– ip address 10.10.10.2 255.255.255.0
– no shut
The following shows the output of the “show vlan id 1” command on the Cisco C3750 switch:
Cisco C3750 - “show vlan id 1” output
From the below, you can see I have no problem pinging 10.10.10.2/24 on the Cisco switch from the Force10 switch.
Pinging 10.10.10.2/24 on the Cisco switch from the Force10 switch
Twitter
LinkedIn
Youtube
RSS
Pingback: Humair’s Blogs » Blog Archive » Trunking and DHCP Across Dell Force10 Switches
Hey,
Thanks for this write up it.
Very interesting.
Would this same scenario apply itself to connecting any two dissimilar switches together?
I would like to port-channel/trunk a Dell Force10 S4810 [8.3.12.0] to a Dell 6248 using fiber.
cheers
Hey,
Thank you for your response.
Your blog was very informative and using your scenario we were able to port-channel our S4810 and our Dell 6248 together without issue.
Thank you again for your examples!
Cheers
No problem; I’m glad you found the post helpful!
Your blog on Force 10 are excellent. I just got several F10 switches and my Cisco experience only got me so far :-(. Your blog helped me get up and running.
One thing I would add to this blog is the Portmode Hybrid.
I ran into one small gotcha!
On the Port-Channel 1 you MUST set portmode hybrid before switchport. You cannot change the portmode if switchport is done first.
I would imagine most people would not have a problem as they would do it the correct way first 🙂
Thanks for help.
Blair.
Blair – glad you found it helpful. FTOS CLI is similar to IOS CLI in many ways, but the way trunking is done is definately one of the main differences. Yes – the ‘portmode hybrid’ command trips up a lot of people if they don’t configure it first; thanks for pointing that out.
Great help, i appreciate the time you dedicated to this post, it was very useful.
Excellent information which helped me partially achieve my objective. Now to ask you to pull the rabbit all the way out of the hat! Combine LACP etherchannel with 802.1q. I can get each one working on its own, but it seems to me that the two elements are mutually exclusive in FTOS! Im using and OLD S25N with FTOS 7.8.1.3. I keep getting “% Error: Port is not in default mode.” when I trie to apply LACP to ports in a tagged etherchannel, or to channel members to ports which are in LACP mode.
It is possible my troubles are rooted in my attempt to combine the commands , “port-channel-protocol lacp” and “channel-member GigabitEthernet 0/1” , the latter which us used to create a STATIC aggregate.
Excellent information, but I have some problems with my Force 10 S50-01-ge-48T having FTOS Ver. 5.8.2.71. Its default vlan is Vlan 1. I am not able to change the default Vlan. By default it also a native vlan and I am unable to tagg it to any port.
I want to basically change a native Vlan to some other Vlan ID and tagg the vlan 1 to a port which is used for trunking with Cisco 3750.
Jiwan,
I am not aware of any 5.8.2.71 FTOS version. Perhaps you are on a really old FTOS version. The last FTOS version released for S50 is 8.4.2.7. Check the version with the “show version” command. Make sure you are not running SFTOS which is a completely different and a very old OS that used to run on old Force10 switches. The last SFTOS code released was version 2.5.3.7. SFTOS is no longer supported. However, it is possible to convert from SFTOS to FTOS. I highly recommend to update the FTOS code.
Hi Steve,
Hopefully, you have resolved this. Although you are running a very old FTOS version (the last version released was 8.4.2.7), there should be no issue here. Sounds like the port is still in layer 2 mode. Remove it from layer 2 mode via ‘no switchport;’ command and then add it to the port channel via static LAG or LACP. Next, go under the port channel and put it in ‘hybrid’ mode and then back in layer 2 mode. The interface should be in default mode (no configuration – it should just have ‘no ip address’ and ‘shutdown’ or ‘no shutdown’ listed under the interface). Below is an example:
”
– enable
– conf
– interface gi 0/1
– no switchport
– port-channel-protocol lacp
– port-channel 1 mode active
– exit
– interface po 1
– portmode hybrid
– switchport
– exit
”
Now you can trunk whatever VLANs you want over the port-channel.
Hi Humair,
Do you know if Force 10 S55 can do inter-vlan routing correctly ? In example, in a Cisco multi layer switches (example packet tracer 3560 24PS), all I need to do is enable “ip routing” command and things just work out of the box when I do ping from vlan 1 to vlan 2.
On a Force 10, this doesn’t seem to be the case for me. I’m using FTOS 8.3.5.3.
Regards,
Aik
Hello Humair,
Thx for the detailed guide. I have a question, if i have a similar scenario with force10 s25n and cisco 4900m inter-connected with a single 10g fiber link. My question is : i iust plugged the fiber cables and all worked fine, how is that possible without any trunking or configuration ? And do i need a trunk in this case ?
Thx
Ali
Doesn’t work on S55 to Cisco 3560. Thankfully we are pulling off anything made by Dell and throwing it in the dumpster. F10, EQL and Sonic Walls all getting replaced with Cisco.
Brian, it will work with S55 just fine. I don’t have the whole context of your issue here, but FTOS (DNOS) OS is the same across the Dell data center switches and this is pretty basic stuff with a standard protocol. Absolutely no reason why it wouldn’t work; I’ve done it many many many times across Dell, Cisco, Juniper, Brocade, etc…. Would be ridiculous if 802.1Q implementation did not work across different vendors.
Also, not sure where the anti-Dell sentiment is coming form here, but I’ve deployed end-to-end Dell solutions with absolutely little to no issues. Many vendors including Dell, Cisco, Juniper, Arista, etc have great products, and, as expected with any product, some bugs as well, but, more often than not, I have found the #1 issue is often design/configuration/implementation and not the product itself.
Now, going all with one vendor for your data center needs can have its benefits but definitely also has its cons, all of which I won’t go into detail here other than saying I wouldn’t do it.
Humair,
Thanks much for this. Helped me immensely! Was killing myself trying to LAG and trunk a N4032 stack to a S25N stack, via 10G fiber. Your initial blog and subsequent reply to Steve saved they day.
Thanks again!