New to HP Procurve switches - 2 questions to start with

[Jon White]

I am a Cisco veteran of Cat 6500, 3750, 2960, etc. etc. I am now working at a company that has Procurve 4108 switches. I have two questions that I hope are easy, but I don't have much time to devote to learning at the moment:

1> I see various ports that have the following reported in the logs:

I 02/27/08 15:14:14 ports: port B1 is now off-line
I 02/27/08 15:14:17 ports: port B1 is Blocked by LACP
I 02/27/08 15:14:20 ports: port B1 is now on-line

Can somebody give me a brief explanation of what this is and just how bad it potentially is? Is it similar to Cisco's 'port is flapping' issue? What is LACP up to in this case?


2> More importantly, I have HP servers that some are using Team'ed NICs. In the switch, the two ports that the wires are plugged to are showing up as:

A3-Dyn5 100/1000T | No Yes Up 1000FDx off
A9-Dyn5 100/1000T | No Yes Up 1000FDx off

To set up proper monitoring on this port(s), do I select one port, both ports or is it even possible to monitor 2 ports that team'ed NICs are using?

Thank you.

[Richard Brodie_1 · Accepted solution]

1. LACP is in passive mode; it's just delaying the link up for a few seconds to wait for incoming LACP packets. It's similar to what would happen if you STP enabled on the port.

There isn't really any significance to the LACP message on the rising edge (offline->online). It's just a single link down/link up event.

[Jon White]

Thank you for confirming my thoughts on what LACP is and what it does.

I am still looking for an answer to #2:


2> More importantly, I have HP servers that some are using Team'ed NICs. In the switch, the two ports that the wires are plugged to are showing up as:

A3-Dyn5 100/1000T | No Yes Up 1000FDx off
A9-Dyn5 100/1000T | No Yes Up 1000FDx off

To set up proper monitoring on this port(s), do I select one port, both ports or is it even possible to monitor 2 ports that team'ed NICs are using?

[Mohieddin Kharnoub]

Hi

Just enable monitoring for both ports and make a mirror port as an output.
I don't think teaming has a big effect on monitoring configuration because the NICs are teamed from the server side, and not trunked from the switch side.

Good Luck !!!

[Chris Stave]

To add a bit to the answer for your first question, you can throw in a

no lacp

for each interface that you aren't using for link aggregation and it will cause the connection to come up faster and keep your logs cleaner.

[Jon White]

The next time that I want to set up Monitoring on two ports that belong to a Teamed NIC of a HP server, I will definitely do it in a test environment first. It still appears that I did nothing wrong. Perhaps it is the Procurve switch that has a problem yet to be revealed.