Intel Boot Agent GE v0.4.09

[Mischa]

Hello,

I'm trying to build a personal supercomputer with four Gigabyte X570S, with embedded 'Intel Corporation Ethernet Controller I225-V' network card, motherboard based 16-core computers. To avoid installing and configuring as many operating systems as I have nodes, I want to use PXE (Intel Boot Agent GE v0.4.09 (X570S) and PXELINUX 6.04 (syslinux.org)).

But, once every so many attempts to boot, I get the following error message:

PXE-E61: Media test failure, check cable

PXE-M0F: Exiting Intel Boot Agent.

Reboot and Select proper Boot device or Insert Boot Media in selected Boot device and press a key

Because there is nothing physically wrong with the cable, a simple reboot suffices to get the system up and running, but I'm really bothered by this 'press a key'. PXELINUX takes care of such problems by rebooting automatically after a number of seconds, such that manual intervention (attaching a monitor, a keyboard, getting a chair etc.) is avoided. It would help if the Intel Boot Agent acted in a similar fashion.

Hope to hear from you soon.

Best regards,

Mischa Baars.

[Caguicla_Intel]

Hello Mischa, 

Thank you for posting in Intel Ethernet Communities. 

Just to confirm, does this mean that you successfully configured PXE boot but the issue is, you are getting the error below once for approximately multiple attempts to boot? 

PXE-E61: Media test failure, check cable

PXE-M0F: Exiting Intel Boot Agent.

Reboot and Select proper Boot device or Insert Boot Media in selected Boot device and press a key

We look forward to your reply. 

Should there be no reply, we will follow up after 3 business days. 

Best regards,

Crisselle C.

Intel® Customer Support

[Mischa]

Hello Criselle,

Exactly, from the screenshots you can see that I've configured one server and four diskless clients. This is exactly the issue I have. One of two to be precise. I will get back to the second.

At the moment, the four diskless clients are Chinese dualcore mini pcs, which will have to be replaced in time. The server is a Gigabyte X570S, with embedded 'Intel Corporation Ethernet Controller I225-V' network card, motherboard based 16-core computer. I would like to see four of these running in parallel, but I experience problems with the Intel Boot Agent.

Hope to hear from you soon.

Best regards,

Mischa Baars.

[Mischa]

I'd like to add:

The Chinese mini pcs are running 'Intel UNDI PXE-2.1 build (0.83)' firmware (on top of 'Realtek Semiconductor Co., Ltd. RTL8111/8168/8411 PXI Express Gigabit Ethernet Controller (rev 0c)' network adapters), and have never shown any problems similar to this one.

From the connections in the screenshot as well, you can see that the cabling itself shouldn't be a problem.

[Caguicla_Intel]

Hello Mischa, 

Thank you for the reply. 

Please allow us to further check your request with our engineers. We will give you an update as soon as possible but no later than 3 business days. 

Hoping for your kind patience. 

Best regards,

Crisselle C.

Intel® Customer Support

[Mischa]

Hello Crisselle,

Thank you for your help

Best regards,

Mischa Baars.

[Caguicla_Intel]

Hello Mischa, 

Appreciate your patience on this matter. 

We already escalated this request to our engineers and they have a clarifications on this one and we hope you don't mind providing more information.

We may need to confirm first if our understanding is correct that the issue only occurs intermittently. Sometimes the Intel Boot Agent does properly work and sometimes it does not. Aside from this, have you tried power cycling the unit (turning off then remove from power outlet) to check if it resolves the issue. 

Additionally, have you also checked with board manufacturer, Gigabyte for the optimal configuration in BIOS for the setup you are trying to achieve.

Looking forward to your reply. 

We will follow up after 3 business days in case we don't hear from you. 

Best regards,

Crisselle C.

Intel® Customer Support

[Mischa]

Hello Crisselle,

The problem indeed occurs intermittently. I have not tried power cycling the unit, nor did I have direct contact with Gigabyte on this specific matter, only on the absence of a BIOS write-protect and on the PCIe version used when cards of different versions are present. I do believe that analog cables are a drawback in digital systems and can sometimes produce unexpected results.

But, as I see all of your responses and your eagerness to confirm all that I am saying, I do think that you are missing the point. Although I don't see similar error messages on other machines on the same network, it is not specifically the error message displayed that is worrying me, it is that fact that the system is waiting for a user to press a key after that.

If there are a number of computers in a server cabinet, with only one monitor and keyboard attached, chances are that we won't be able to see these error messages and it is very difficult to reach a specific machine when an error occurs. It would be better, as I tried to make clear in my first message, if the machine reboots by itself after a certain number of seconds and tries again. There is nothing wrong with the cables, as far as I'm concerned, because TCP/IP is working. The waiting for the user to press a key is the problem and the primary issue at hand.

Hope to hear from you soon,

Best regards,

Mischa Baars.

[Caguicla_Intel]

Hello Mischa, 

Thank you for the swift and detailed response.

Please allow us to re-escalate this request to our engineers. We will give you an update as soon as we heard from them but no later than 3 business days.

Hoping for your kind patience. 

Best regards,

Crisselle C.

Intel® Customer Support

[Mischa]

Hello Crisselle,

Thank you for your help!

Waiting for your reply,

Best regards,

Mischa Baars.

[Mischa]

Hi Crisselle,

Are you still there?

The actual error might be triggered because the firmware measures the wrong internal wire resistance, if it measures such a thing. I'm using CAT5e cabling, instead of CAT6 or CAT7 cabling.

Best regards,

Mischa Baars.

[Caguicla_Intel]

Hello Mischa, 

Please accept our apologies for the delay on this matter. 

Be advised that this request is currently escalated to our engineering team. We are still waiting for their feedback and rest assured that we will give you an update as soon as we heard from them. 

We will also forward your latest update on this thread and get back to you no later than 3 business days. 

Hoping for your kind patience. 

Best regards,

Crisselle C.

Intel® Customer Support

[Mischa]

Hello Crisselle,

According to Logilink, supplier of network cables, CAT5e cables do support 2500baseT/Full. The DHCP server and the switch in between are 1000baseT/Full. The resultant connection therefore is 1000baseT/Full. All cables are CAT5e minimum. This connection has been tested under Linux using the TCP/IP protocol from 512byte upto 65kbyte data fragments both half and full duplex.

It should not be the cables, which makes both escalations, the actual error message and the automatic reboot, equally important.

Best regards,

Mischa Baars.

[Mischa]

Hello Crisselle,

Yesterdey, I tested the cables with a piece of my own software. About 50 Gbyte of data was transmitted to the second computer during 1024 trials (1024 connects and 1024 disconnects). Not a single transmission error.

This makes it very unlikely that one of the cables is not correctly attached.

Best regards,

Mischa Baars.

[Caguicla_Intel]

Hello Mischa, 

Good day!

We'd like to inform you that we are still actively working on this issue with our higher level Engineers. Rest assured that we are doing our best to sort this out as soon as possible and we will give you an update no later 3 business days for the status of your request. 

We will also forward your additional comment on this thread to our engineering team.

Best regards,

Crisselle C.

Intel® Customer Support

[Mischa]

Hello Crisselle,

Thank you for your help...

Best regards,

Mischa Baars.

[Caguicla_Intel]

Hello Mischa, 

You're welcome. Thank you as well for conducting additional tests on your end, we will forward your feedback to our engineers. 

We would also like to sincerely apologize for the delay on this matter as we are still awaiting for the feedback from our engineers. Please give us more time to look into this. Rest assured that we are closely monitoring this request and will update you as soon as there is any findings but not later than 3 business days.

Thank you for your kind cooperation.

Best regards,

Crisselle C.

Intel® Customer Support 

[Caguicla_Intel]

Hello Mischa, 

Thank you for the patience on this matter. 

Our engineering team would like to gather below details and we hope you don't mind sharing this with us. 

1. What is your platform current NVM version?

2. What is the method of PXE booting that you are using, over UEFI or legacy?

3. Is there any switch being use between your platform and PXE server? If yes, please remove the switch and retry.

4. What is the cable length that you are using for your environment.

Awaiting to hear from you soon.

Should there be no response from you, I’ll make sure to reach out after 3 business days. 

Best regards,

Crisselle C.

Intel® Customer Support

[Mischa]

Hello Crisselle,

I accidentally considered the two issues equally important. My mistake. On closer inspection, the issue that arises from the cable issue, and that makes the computer wait for user input is so much more important. Perhaps we should focus on it.

1) I am not a Java programmer, but even if I was, I would still not see what Java has to do with PXE booting.

3) Do you expect me to use a cross-over cable instead?

4) Not much longer than the one used to flawlessly transfer the 64Gbyte in 1024 connects and disconnects.

2) The default settings. Equally unimportant, because it boots nine out of ten times. UEFI if I'm correct.

An automatic reboot in case this error occurs would really solve most of my problems. Hope Intel can do that for me.

Best regards,

Mischa Baars.

[Caguicla_Intel]

Hello Mischa, 

Thank you for the swift response. 

Please allow us to forward your reply to our engineers. We will get back to you once we heard update from them but no later than 3 business days. 

Thank you for your kind cooperation. 

Best regards,

Crisselle C.

Intel® Customer Support

[Caguicla_Intel]

Hello Mischa,

Thank you for the patience.

Please see below feedback from our engineers and we hope you don't mind providing the additional information they request. 

LAN on motherboard (LOM) such as Ethernet Controller I225-V has its own NVM/Firmware and our engineers would like to capture it for further validation. Is it possible for you to check with the motherboard manufacturer what is the default NVM version of your Ethernet Controller?

Awaiting to your reply. 

We will follow up after 3 business days in case we don't hear from you. 

Best regards,

Crisselle C.

Intel® Customer Support