LACP and ESX 5.5

[Steve]

Setting up link aggregation between a NRS 5520 and our ESX Servers.

Used Nortel configuration commands:
interface FastEthernet x,y,z
lacp key port x,y,z <number>
lacp mode port x,y,z active
lacp aggregation port x,y,z enable

And carried out the suggested configuration on my ESX Host.

Probably being a twit, but I can't find a show command in the CLI reference guide, to see if my port channel is up. With Cisco I'd use something like 'show etherchannel summary' to see the PAgP equivalent.

Can see all the configuration is as expected in a show run and connectivity OK so assuming all OK.

Regards,
Steve

[Dominik]

Use this command:

show mlt

An LACP trunk should be show uo in the mode DynLAG

Cheers

[Steve]

Hi Dominik,

Thanks for the reply.

No, not seeing mode DynLAG. Seeing 1-32 but no MLT members and mode is Basic for all.

I was actually looking to use LACP/802.3ad (as this required for ESX), rather than MLT. I believed they were alternatives on the Nortel platform?

Am I missing the point somewhere?

Regards,
Steve

[Dominik]

It looks like your LACP trunk is not running.

Check this blog post from Michael, here you find a good manual, how a LACP trunk is setup on a
Avaya switch.

http://blog.michaelfmcnamara.com/2009/08/lacp-configuration-examples-part1/

It is a littlebit confusing, but to show the state of an LACP trunk you have to use the command sho mlt or sho mlt info.

For more details, take a look at this Avaya guide, it shows in detail how to connect an ESX host to an Avaya switch:
https://www.google.de/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&ved=0CCEQFjAA&url=https%3A%2F%2Fdownloads.avaya.com%2Fcss%2FP8%2Fdocuments%2F100123707&ei=4G1rVOejHsTNOJmlgPAN&usg=AFQjCNGfzuZig9nd-dLZwAa3Ksb84LQTCg&sig2=poiuo3SamXGmI6LBZ4QykQ&bvm=bv.79908130,d.ZWU

[Steve]

Hi Dominik,

Thanks for the info, I hadn't seen the Avaya ESX recommendation document, so much obliged for that.

I had been using Michaels blog to create the port channel though. I wasn't entirely looking to do trunking across the port channel, really just looking to do LACP. I know LACP and tagging are usually carried out in tandem but they are exclusive of each other and can be used separately. Hence just doing just the link aggregation part. Reason being, I am only carrying one VLANs traffic between the 5520/ESX Servers, so no real reason to tag frames.

I hadn't clocked the 'show lacp ports' command in Michaels blog however. So have run that and being told my LACP status as active.

I've attached a screendump of the output of the command, I'd say my LACP is active would you not?


Thanks again,
Steve

[Steve]

Oh and ignore the fact that my ports are on the same unit. I've distributed them over different units in the stack now.

[Steve]

Solved!

Missing a little config on the ESX side.

Thanks
Steve

[Dominik]

Thankks, Steve for the feedback.
Good to hear that it finaly works.

What did you miss on the ESX side ?

Cheers

[Steve]

Good shout, should have reported my fix.

Basically I hadn't set up a LAG in my distributed vSwitch. I was trying to do load balancing directly on the port group. Similar to how it was in done in 5.1. Created the LAG, assigned to the physical NIC and working now.

Should also add, I am doing this without tagging VLANs.
Just adding for posterity. If anyone comes to this at a later date and has the same hesitations I had.

Steve