NIC Teaming issue Windows Server 2019

Question

Good day,

2019 server, 4 NICs. (dell R710, QLogic BCM5709C nics x 4, DLink managed switch)

Have been using NIC teaming for many years with success on a 2012R2 similar hardware system, one generation older, and while I initially had some of the "network startup woes" where it was fine until a reboot, but then team wouldn't come online... we resolved that fairly quickly, adding a dependancy somewhere along the way, as I recall

This new system, I have been running as a lab machine, but need the bandwidth, and I'd really like it to work too!

I have been playing with a few Unix VM systems, the last of which was ProxMox, which took the LACP config like a pro, and I bonded the 4 NICs to one IP, plugged them in the 4 ports that had been configured to do 802.3ad (LACP), and boom, 4gig connection worked like a champ.

Have installed a headless 2019, to run the HyperV on, and while initial NIC config was direct and automatic, once I created the Team and added the 4 nics to it, and moved the cables over to the 4 ports they should be in, no joy.

I get an UP status for the 4 ports, or disconnect if I pull them.

The team originally said "cable disconnected", but now that I have sent a few enable/disable commands to try and "tickle" it, it's status is just "blank".

My team config is set to: includ the 4 names, lacp mode, dynamic, and fast.
(those happen also to be the same settings the old 2012r2 has been running on for years, except "fast" as that seems to be a new setting since 2012r2)

Doing this via powershell now of course, at the console, so open to any commands, tweaks, etc.
The only other thing i know that ProxMox was setting, was the hash type, and we were using the default of level 2/3.

Open to any thoughts or ideas, or things I can try. :-)

with thanks,

Andrew

Accepted Answer

Solution:

Moved the physical box into a live room, where there is a Cisco SG220-50 smart switch, presently doing NIC teaming with a 2012R2 box for a number of years now.
Created a second LACP trunk, assigned some ports, and ran haphazard cables from the server box to this switch.

Bam.
Success, with two ports, then three, now all 4 as I wanted it.

It would seem that Windows Server 2019 (and perhaps all of them before it too?) are sensitve to the switching hardware - where other operating systems are not.
(recalling that I had LACP working on this same box, with both the DLink smart switch mentioned above, and an SMC smart switch that I dragged out to the lab - using Debian Linux, but neither switch liked Windows server teaming)

:-)
Andrew

Answer

Hi ,

The team originally said "cable disconnected", but now that I have sent a few enable/disable commands to try and "tickle" it, it's status is just "blank".

Could you upload some screenshots to help us understand your issue better?

For how to set up NIC Teaming in Windows Server 2019, you can refer to the following link:

https://techsnips.io/snips/how-to-set-up-nic-teaming-in-windows-server-2019/

Best Regards,

Candy

--------------------------------------------------------------

If the Answer is helpful, please click "Accept Answer" and upvote it.

Note: Please follow the steps in our documentation to enable e-mail notifications if you want to receive the related email notification for this thread.

Answer

Thanks Candy, your video link helped me to find the newest GUI options, and make this process much simpler. As a headless server, I had been working through powershell, where a gui lets you flip things back and forth and try ready examples, for sure!! :-)

I have also since original posting, changed out the drivers for the NICs, for drivers from Dell, specifically for these network cards listed above.
Not sure if that helps or hinders, happy to listen to advice on that. There was commentary from others (elsewhere) that suggested the MS default drivers for these cards has "issues".
But that didn't help either.

I can confirm that the switch is set for 4 ports of LACP, and I am presently using three of them, while keeping the 4th nic separate from the team, to continue to have remote access to the server.
Switch is a DLink, and the ports are set to "active". (that was raised elsewhere as a passive/active setting choice, their default is active)
Speed on the switch for ports, on a timeout, is set to "long (90s)", where the alternate option is short (3s).

Creating and changing the team in the gui, is working, and is resulting in a pause, and then a "Faulted LACP Negotiation".

one Event recorded is, but not every time:

16947 x 3 (one for each nic)

LACP Churn detected on Member Nic {e1bccf97-6dca-4ef5-bf6a-f1dea38bcd2a}.

So, maybe the switch is an issue - as in, windows is asking for something it can't do?
Yet strange, that Debian Unix asked, and got it...first try?

The Debian setup asked for "hash type", and I chose "Level 2+3", for reference sake. I didn't try anything else, because that worked.
But "hash" option on this server under windows teaming, also failed.

Thanks for the ongoing support and ideas! :-)
(and, cool handle by the way)
Andy

Answer

Anyone else run into this at all?
I've been experimenting with the headless server the last two days now and am really enjoying it, but if I can't figure out how to have a 4gb channel to/from it, I'm sort of screwed for using it. :-(

Answer

Hi ，

Creating and changing the team in the gui, is working, and is resulting in a pause, and then a "Faulted LACP Negotiation".

Based on my research, it seems that Symantec client running on the server will cause an issue with getting LACP negotiated.

Have you installed any third-party AV software? Remove any third-party AV software and check again.