Solved: WARNING: Link is up but PHY type 0x3 is not recognized – Can cause ESXi 6 purple screens
The Error >> When running an Intel x710 NIC with the ESXi i40e driver you notice your vmkernel.log completely full of the error “WARNING: Link is up but PHY type 0x3 is not recognized”
The Solution >> Ensure X710 firmware is at 17.5.11 (aka 5.04 in ESXi) and ESXi i40e Driver to 1.4.26 or 1.4.28 and these errors stop
The Follow-up >> Check out your NIC on the VMWare HCL for the Correct driver/firmware guidance. This is the link I used.
Other notes…
Sending Millions of the PHY errors to your event logs could be causing other issues for your ESXi host. Look for local boot disk latency or Networking errors in your ESXi host event logs. Once you apply this solution these issues should stop. If not, then you may have other issues impacting your boot disks.
*Updates*
- After applying this solution we then noticed the vmkernel started to populate with ‘driver issue detected, PF reset issued’ the solution for this is to disable TSO/LRO. VMWare KB 205140.
- 04-10-2017 There is a new VMware driver listed for the X710, will be testing soon and will post up results. Release notes indicate fixes for the following:
– Fix duplicate mulicast packet issue
– Fix PSOD caused by small TSO segmentation
If you like my ‘no-nonsense’ blog articles that get straight to the point… then post a comment or let me know… Else, I’ll start writing boring blog content.
vmexplorer 
February 10, 2017 at 7:43 am
I hope this fixes the issue! VMware still has this listed as a known issue, but I ran across this and gave it a try last night.
LikeLike
February 10, 2017 at 8:29 am
Thanks for the comment. You mentioned VMware still lists this as an issue. Can you please send me the context (KB, URL, etc.) that you are referring to? I’ll be glad to look at it. My advice on the X710 is, keep that firmware and drivers up to date. If you are current on your VMware licensing you might check out Log Insight, its free up to 25 objects and its the tool I used to find and solve these issues.
LikeLike
February 10, 2017 at 8:49 am
Here is the URL to the KB – https://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=2126909
The error message you speak of isn’t mentioned there, but this is the KB I was sent when I was having that issue.
LikeLike
February 10, 2017 at 8:54 am
Thank you for the quick response. The KB you mentioned (KB 2126909) is correct, this still is an issue for the x710. Without this workaround you might notice PF Reset errors in your logs and PSOD issues. To remove both issues (1 – PHY and 2 – PFResets) I updated my firmware/drivers and applied this KB. From there we have been rock stable, no issues. I hope this helps, and if you are still having this issue, I’d recommend opening a VMware service request and have them look at your system.
LikeLike
July 10, 2017 at 10:27 am
How is the April release of the driver? Are you still seeing errors or have you had any problems?
LikeLike
July 10, 2017 at 10:32 am
Hey Matt – We should be testing this soon. I’ll post up if I find out whether or not disabling Off-loading will be needed with the new driver.
LikeLike
January 30, 2018 at 11:12 am
Hello Matt,
Has there been any update to this? I’m wondering if this is still an issue, or whether or not it was solved in the latest driver updates. Thanks!
LikeLiked by 1 person
February 1, 2018 at 2:03 pm
Hey Mitchell — This item got placed on the back log so I have not been able to test if enabling TSO/LRO has any impact. What I would recommend is update the driver and firmware and if the vmkernel.log displays the error “WARNING: Link is up but PHY type 0x3 is not recognized” then you’ll need to keep TSO/LRO disabled.
LikeLike
February 1, 2018 at 2:27 pm
I found this article last year and updated to the latest firmware & driver in March and enabled TSO/LRO on my 3 esxi hosts. No errors pop up in the log and no purple screens since then. The driver version I have installed is 1.4.28 and firmware 17.5.11.
LikeLike
February 1, 2018 at 3:55 pm
I am running the driver version 1.4.28 and firmware 17.5.11.
I will be updating one of my hosts to the newest driver you mentioned (2.0.6) tonight. I will post back in a few days to let you know if it is still error free.
LikeLiked by 1 person
February 8, 2018 at 9:32 am
I upgraded the driver to 2.0.6 on 2/1/18 on my hosts, and I am still error free with no purple screens.
LikeLike
February 8, 2018 at 9:44 am
Ariel — Thank you for posting up.
RE: I upgraded the driver to 2.0.6 on 2/1/18 on my hosts, and I am still error free with no purple screens.
Questions:
What NIC firmware are you running?
Is TSO/LRO enabled?
In the vmkernel log do you notice any ‘driver issue detected, PF reset issued’ or ‘PHY type 0x3’ notifications?
LikeLike
February 8, 2018 at 9:52 am
The firmware version on the updated hosts is 17.5.11. I have TSO/LRO enabled. There are no notifications for any of those things in the vmkernel log for the updated hosts.
LikeLike