Maj bzimage

[Sebastian Roth]

@jps Ok, then you don’t need to worry about the message.

[jps]

It is still necessary to update the file ?

[Sebastian Roth]

@jps You mean the kernel? Not as long as imaging works.

[jps]

It only works at 100 Mbits.

[Sebastian Roth]

@jps said in Maj bzimage:

It only works at 100 Mbits.

This is the first time you mention this.

When you reboot the machine to it’s installed OS, does it run on gigabit then? Just want to make sure we are looking at a driver problem here.

[jps]

When I test with the 5.6.18 it works at 1Gb.

[Sebastian Roth]

@jps Then I’d say you better use the “unofficial” 5.6.18 kernel.

[jps]

When I use the “unofficial” 5.6.18 kernel I can’t connect the machine Dell 5500/5300 in 100 MB or 1 GB.

[george1421]

@jps said in Maj bzimage:

“unofficial” 5.6.18

Just so I’m clear, what do we mean unofficial? Is this the one that is downloaded from the FOG servers using the built in FOG web ui or my one-off kernel?

For these Dell 5500’s can you provide the hardware IDs of this network adapter? It will be in device manager -> device -> details -> hardware IDs. The value we need looks like this PCI\VEN_8086&DEV_15BB&CC_0200 Also how are these network adapters connect to the computer? Are they built in, usb dongle, WD19 port replicator, how?

Lastly what do you mean you can’t connect? Do you see the FOG iPXE menu, or you can see the menu but when you pick an option FOS Linux can’t see the network adapter?

[jps]

I repeat what Sebastian says “unoffical”, I mean “Kernel.TomElliott.5.6.18.64” . The hardware IDs is PCI\VEN_8086&DEV_15BE&CC_0200 built in. When I pick an option FOS Linux can’t get an adresse IP from DHCP and don’t know if FOS see the network adapter.

[george1421]

@jps The network adapter 8086:15BE first appeared in the linux kernel 4.12 and exists in these kernels [ 4.12–4.20, 5.0–5.8 ]. With that said 5.6.18 should have the required driver installed. (it should just work).

To debug this issue schedule another deploy/capture task but before you hit the schedule task button tick the debug checkbox then schedule the task.

PXE boot the 5300 or 5500 computer. After a few screens of text you need to clear with the enter key you will be dropped to a linux command prompt on the target computer. This is the FOS Linux command prompt.

Please post the results of the following commands. (a clear screen shot of the text taken with a mobile phone will work here).

lspci -nn | grep -i net

ip a s

uname -a

Lets start with those. I have another command for you to key in but lets make sure we are getting an output of those first.

[jps]

This post is deleted!

[jps]

Here they are :

[george1421]

@jps Ok looking at the pictures everything looks good (it should be working), but do see for eno2 there is no carrier (cable not plugged in). I am a little surprised as its named eno2 but if that is what the hardware selected its ok, because its a logical name.

Do we know why its saying the network cable isn’t plugged in? I would think that it had to be if the system pxe booted this far.

If you are still at the fos linux command prompt you can test this command and see what it produces
/sbin/udhcpc -i eno2 --now

[jps]

Here it is
I specify I re-run “ip a s” and there is no “no-carrier”

[george1421]

@jps Ok this is interesting an also makes me think.

I want you to do this in sequence step by step. If you are still at the command prompt just reboot the target computer the job will still exist in FOG.

1. PXE boot the computer in debug mode.
2. As soon as you get to the FOS Linux command prompt key in ip a s. Confirm that you do not have an IP address
3. Wait ~30 seconds.
4. Key in /sbin/udhcpc -i eno2 --now
5. Confirm that it detects an IP address.

If the above does work then time resolves the problem. This is usually an indication of spanning tree being used on your building network switch. I want you to check with your network team to ensure on of the fast spanning tree protocols (portfast, RSTP, MSTP, fast-stp, or what ever your switch mfg calls it) is used on the network port where the computer is. Standard spanning tree reacts to slowly because FOS Linux boots in 10 seconds and standard spanning tree doesn’t start forwarding data for 27 seconds. By the time standard spanning tree starts forwarding data FOS Linux has already given up. That is what I think.

[jps]

I could not reiterate the case where I have no ip address. On the other hand at boot (not as a debug task) I have this:

[jps]

This post is deleted!

[jps]

@Sebastian-Roth When I reboot the machine to it’s installed OS it runs on gigabit.

[Sebastian Roth]

@jps command not found…?! Possibly and issue in the scripts?! I will take a look.