1

Question: To setup uplink ports from network switch in trunk mode vs putting them in a port-channel?

We use Dell N series and PowerConnect switches in our agency. We are installing a new rack for our servers. Currently we are specifying the uplink ports to be part of a port-channel back to our core switch. If there is only (currently) one uplink connection is it better to leave the ports in the port-channel or just specify them as trunk ports?

interface Te1/0/1
channel-group 1 mode active
description "Uplink to Core"
exit
interface port-channel 1
classofservice trust ip-dscp
switchport mode trunk
switchport trunk native vlan 10
exit

vs

interface Te1/0/1
classofservice trust ip-dscp
switchport mode trunk
switchport trunk native vlan 10
description "Uplink to Core"
exit

Would initially setting up the switches with the port-channel allow for easier expansion in the future without interruptions if we decided we needed more bandwidth?

Update: Decided to go with LACP between the the switches just in case we needed to add more bandwidth in the future. This rack that I am setting up is for a healthcare agency. There will be iSCSI traffic going across these switches as well as additional user traffic.

  • This is a fairly simple question, but one only you can truly answer. If you forecast increasing bandwidth, then do it now. If it's just a possibility, make a judgement call on whether you can part with whatever petty resources it's going to cost. – Ryan Foley Sep 30 '14 at 18:18
1

I would suggest using Port-channel if you intend to add more links in near future. You have enabled LACP which will take care of additional links once added.

HTH

2

My general rule of thumb: if you don't need or aren't currently using a feature, don't enable it.

Enabling features you don't need/use only exposes yourself to more problems. At the very least, this adds unnecessary lines to the configuration, making it longer and/or more complex. The longer and more complex a configuration, the more prone it is to human errors and it can slow down troubleshooting processes.

It can also expose you to bugs/problems in the code that you would not otherwise be subject to without the unnecessary service.

If this is something you plan to implement in the short term, then you may consider it. However keep in mind that when you are adding your additional links, you can add the link aggregation at that time with a minimum of downtime (if done correctly).

  • 1
    True. But when he wants to go to port-channels, his network will take a hit during the transition. – Ricky Beam Sep 30 '14 at 20:14
  • 1
    @RickyBeam, as I said in the last paragraph, the hit should be minimal if done correctly. (Correctly includes that they are using RST/MST.) The second link can be configured for the link agg and brought up. Once ST settles in, you bring down the first link, add it to the link agg and bring it back up. There should only be the one ST transition. – YLearn Sep 30 '14 at 20:21
  • 1
    Short answer: the network will take a hit We don't know the specifics of his business. Maybe a 1hr downtime won't be questioned. On the other hand, a 5s glitch may be unacceptable. – Ricky Beam Sep 30 '14 at 20:42
  • 1
    And again, the last paragraph clearly indicates that there is a downtime (i.e. you can't minimize something that doesn't exist), so I don't know what your issues is here by continuing to insist "the network will take a hit." However, while we don't know the specifics of the business, I will go out on a limb and state that if it is in a "5s is unacceptable" environment, then they better have a design that isn't dependent on a single link (whether that is a single link or link aggregate). – YLearn Sep 30 '14 at 22:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.