The case of the Intel NIC that refuses to work.

[Ampera]

So I used to have a Z97 PC Mate board by MSI. It worked perfectly fine until it didn't. The board randomly fried and in a fit of stupidity I decided that logically the only way to install an LGA cover was like a CPU, thus mangling the socket, and forcing me to buy a new board.

Now I have a Gigabyte board (GA-Z97-D3H) which works great, and is even better than my previous board on the count of the M.2 slots (which are unused but nice for the future). My issue is that regardless of what I try the integrated network adapter will not work, at least not fully.

Originally what I tried to do was install the driver from Gigabyte's website. It claimed I didn't have any Intel NICs on my computer. This is bullshit, as it says right on the box it has that adapter, and it's from the driver download page for the motherboard. The only variable in play here is that I am using Windows Server 2016, but this shouldn't mater as it should take signed Windows 10 drivers, and at the very least give some sort of error for the driver being unsigned.

However, this doesn't happen. Even pointing Windows to the folder with all the drivers in it, it refuses to install any of them.

The card is an Intel I217-V according to looking up the hardware IDs. What I did do was manually install the driver for the Intel I210 Network Connection, which sort of works, but the only issue that it refuses a manually assigned IP.

What I mean by that is if you attempt to use ANYTHING but pure DHCP automatic IP assignment, it doesn't get working network access. If I put a manual IP assignment in Windows, it does this, but it also does this if I go into my router and put in a DHCP reservation.

I have never had it happen for a NIC on a mobo outright refuse to work right. Every single NIC I have ever used worked perfectly fine, with minimal exceptions, so I am completely confused here. Any help would be great.

[amyk]

That's really an uppercase I, not a lower case 'l' nor the number '1'.

Horrible product name. They could've used a lowercase 'i' or an uppercase 'L' (for LAN, as a lot of people seem to assume) and made the name a bit less confusing.

[Halcyon]

The NIC controller chip should be visible on the board near the NIC ports. Have you tried to identify the chip set by physically looking for it?

[Ampera]

No, because there is a massive heatsink I do not wish to remove from all the chips near the RJ45 jack.

And yeah, it's an uppercase I. I was confused by this too, as an upper case I can look like a lowercase l in some fonts, including the one used on the Windows management console.

I already identified the chip, both from Gigabyte's website and through Hardware IDs. It IS an I217-V, and there can't be any doubt about it. The fact that the I210 driver works adds a bit more merit to that. It just refuses to install from any drivers, and I do not think it's because of a signature issue.

[Jwalling]

Did you see this page?
https://blog.workinghardinit.work/2017/06/19/installing-intel-i211-i217v-i218v-i219v-drivers-windows-server-2016-eufi-boot/

[Ampera]

Did you see this page?
https://blog.workinghardinit.work/2017/06/19/installing-intel-i211-i217v-i218v-i219v-drivers-windows-server-2016-eufi-boot/

Thanks for that. No I did not see it, and I am not entirely sure how. My Google-fu skills need work. At this point I don't even want to question how it works. Manual IP assignments also work too.

[Jwalling]

Thanks for that. No I did not see it, and I am not entirely sure how. My Google-fu skills need work. At this point I don't even want to question how it works. Manual IP assignments also work too.

It looks like a fairly involved process. I'd love to hear the outcome, if you give it a go.
Intel makes weird (?) decisions about server O/S support. They decide at or before roll-out whether a chip will be supported in server O/Ss or not. Typically, the NICs and graphics chips are affected. BTDT with graphics support. Perhaps it's due to the increased testing/certifications involved?

[Ampera]

Thanks for that. No I did not see it, and I am not entirely sure how. My Google-fu skills need work. At this point I don't even want to question how it works. Manual IP assignments also work too.

It looks like a fairly involved process. I'd love to hear the outcome, if you give it a go.
Intel makes weird (?) decisions about server O/S support. They decide at or before roll-out whether a chip will be supported in server O/Ss or not. Typically, the NICs and graphics chips are affected. BTDT with graphics support. Perhaps it's due to the increased testing/certifications involved?

It worked quite well. I guess I wasn't clear in saying that I did in fact try it out. The only issue is that Chrome has now decided to run stupidly slow, but I haven't proven that to be a correlation.