trunking

[Young Yu]

I am having some issues with trunking on a 5412xl. I have STP turned off and no lacp. When I configure 2 ports to be trunked manually and do the same to various switches (all HP ranging from 2600 to 2800 series), I seem to get drop outs on multiple servers on the 5412. I've talked to three HP techs just to verify I am not doing anything wrong and was assured this can be done live and would not affect any other ports. Can anyone shed any insight? Thanks in advance.

[Young Yu]

I am currently using revision K.11.41 and know there is a new version out but I can't reboot the unit currently. Is there a known issue with trunking and K.11.41?

[Jonathan Axford]

Hi Young Yu,

What are you trying to trunk between? Is it the 5400xl and the servers or is it between the 5400xl and other HP switches..??

[Young Yu]

I am trunking between various HP switches. Thanks.

[Jonathan Axford]

Hi Young Yu,

I am not sure about problems with that release. When i installed 5 5400xl's about a month ago, our HP rep sent me the latest version of code and insisted that i installed that before configuring the switches and making them live.

It will probably be worth while scheduling in some time to update the versions. I have had no problems with trunknig between 5400 and 2600 and 2800.

Is it definately the trunked links that are dropping traffic? Is it possible it could be the links to the servers themselves?

[Young Yu]

From the logs I can see excessive traffic on ports that aren't even trunked. I've tried this three times now. I think the best thing for me at this point would be to upgrade the revision and then try it again. I really think the issue is specific to the 5412 since I don't have the same issues with trunking between other model HP switches.

[Young Yu]

This morning I up dated to revision K.11.69 on the 5412zl switch and still the same issue. No lacp or stp and as soon as I plug in the first line (out of three) I get a drop from excesive traffic. Am I doing something wrong?

1. Configure trk1 on 3 ports (tried lacp as well as trunk).
2. I do the same for the other switch (tried setting it 5412/trunk-2900/second switch trunk as well as 5412/trunk-lacp/2900).
3. Plug the ethernet cables in (they were disconnected during setup).

Can someone let me know if you see something I am doing that could potentially be causing my issues? I've created trunks between 3 other switches to see if it's the steps but I had no issues with that setup. Is it possible the excessive traffic is being generated by servers when I create the trunk?

[Matt Hobbs]

Is it the port counters or warning messages in the switch log that you are concerned about? Quite often these are false positives.

Are you seeing a negative impact on your network performance?

[Young Yu]

I am getting the following error message in the logs.

"High collision or drop rate on port" followed by a every port number. It's not just a message but network traffic stops or just crawls for the duration. I can't leave it like this for any duration to see if the traffic will subside or even if anything is accessible since this is a 24/7 environment.

At this point I am thinking of getting another 5412 just to see if it's something I am doing. That's a expensive test though.

[Matt Hobbs]

It sounds very much like a loop is being created. Try enable spanning-tree and see if you still get the same errors.

You've verified that the ports are definitely connecting to the trunk configured ports on the other switch?

[Young Yu]

Yes I am definitely plugging into the correct ports. I am thinking there has to be an existing connection between the two switches. The results are typical of a loop. I don't know why I didn't think of it before (such a noob mistake, even though I am fully aware of this). I just have to trace tons of cables to see where they are leading. Thank you very much for bringing to my attention.