Posts made by george1421

@keith75140 Well then stop doing it!!

Can you log into the linux console of the fog server as root or another user account where you can fix things?

Do you happen to use the linux user fog’s account for system management? That linux user fog is a service account owned and managed by FOG. It should not be used for system administration. Is this what happened?

@tom If you setup wireshark on the same subnet as the 7290, you can tell. If you use this capture filter port 67 or port 68 then pxe boot the target computer. Your answer will be in the first discover packet. Just make sure the mac address matches your 7290. But I would setup for both uefi BC (7) and X64 (9) just to be sure.

@flipwalker As Tom said, only a single unattend.xml file is used. If you look at my sysprep command in the link below, there is only one option to provide a path you can provide to the unattend.xml file. Save your self some grief too, place the unattend.xml file in C:\Windows\Panther for windows 10.

And to add on to Tom’s comment about AIK, you need to get/use the current version of the OS you are deploying. There are typically updates to AIK with each new release of Win10.

@tom-elliott I need to look at the document to make sure its not capable of the zombie attach that impacted github, but here is what I setup before: https://forums.fogproject.org/topic/10717/can-php-fpm-make-fog-web-gui-fast/3

[edit] nevermind I already put in the bind to 127.0.0.1 in the tut so we are protected already [/edit]

@tom-elliott Did you also switch over to memcached? or just phpfpm? I’m more curios for performance improvements than the OPs issue. (Oh did I say that on the outside…)

@flipwalker who’s bootable USB?

Maybe you need to check into/for a firmware update?

First I would go throught this article and setup bios/uefi coexistence to keep you from going insane having to deal with both platforms: https://wiki.fogproject.org/wiki/index.php?title=BIOS_and_UEFI_Co-Existence#Using_Windows_Server_2012_.28R1_and_later.29_DHCP_Policy

Bitlocker is a problem that you must address on the reference computer. You must completely disable bitlocker and unencrypt the drive. FOG will copy it encrypted but it won’t ever boot on the target computer because the TPM chip will not be configured correctly. Always enable bitlocker on the target system after fog has done its business.

FWIW: This is a pretty good online generator for your unattend.xml file. Just don’t include any private information (like license keys) in the online form. Add that data afterwards for security reasons.

Here is a post that shows my call to unattend.xml: https://forums.fogproject.org/topic/11542/super-simple-unattend-customize-xml-for-windows-10-1709

Also look at the chat icon in the FOG Forum tool tray for another few hints.

@tom undionly.kpxe is the default. You can use either one interchangeably. In rare conditions where undionly.kpxe doesn’t work, the .kkpxe adds some additional code to band-aid some broken hardware. I can say in my environment I only use undionly.kpxe without issue.

@tom That picture does give me a clue. Where the heck did kernel bzImage4123 come from? Lets reset that back to just bzImage to ensure you have the latest kernel that was delivered with 1.5.0. If your kernel does not match the inits, I can see why FOS did not startup.

@tom said in Issue with FOG 1.5:

1.Is the 7270 in uefi or bios mode? Yes

That was an either or question, but based on your answer to #3, I might guess you are in legacy (bios mode).

For the boot file, can you test to see if undionly.kpxe (note only one k) works better.

@callumhird boy you have quite a wish list. Lets see if we can address them in order. Based on your questions I’m suspecting you are coming from a pretty old version of fog (i.e. 0.3x)??

1. FOG no longer needs FOGprep. Just use your normal sysprep command and have sysprep power off the computer to ensure the disk is closed properly. If you just use shutdown partclone will complain that the disk was not shutdown properly and refuse to upload.
2. This one is a little complicated. (this part may be a bit off-loint) The older versions of fog allowed you to have custom naming rules when you picked quick registration. That code has been removed since FOG 1.1.0. If you go through manual registration you can enter the asset tag there. What we have for our system naming is a LUM code and then the dell asset tag. When we do registration we will key in the LUM code and then use a usb bar code reader to “zap” the barcode on the dell computer’s chassis. This keeps us from having a type-o when entering the system name into FOG. Now once that is done, you can have FOG set the target computer system name. This is done with the FOG client. So the fog client service will need to be installed in your golden image if you plan on using some of the advanced fog features.
3. The FOG client can join the computer to AD using the OU you define in the webgui. In my case we have a complex OU arrangement so the OU is calculated during imaging based on the hardware type (desktop/laptop/tablet), site code, and image name being deployed. This calculated value is then updated in the unatend.xml file dynamically during imaging using a FOG post install script.
4. Use a FOG post install script I have a tutorial on how to do this. I will link it in a few minutes. [edit] Well it looks like the tutorial is all done except for the finishing: https://forums.fogproject.org/topic/11126/using-fog-postinstall-scripts-for-windows-driver-injection-2017-ed
5. The creating a base image is outside of the scope of FOG. It is up to the IT admin on the best way to setup the golden image for their company. In my case I use MDT to build a consistent golden/reference image every time. Then I capture and deploy with FOG.

Lets collect some additional information for this issue.

1. Is the 7270 in uefi or bios mode?
2. Have you updated the firmware in those systems from what was shipped from Dell?
3. What (exactly) do you have for dhcp option 67 {boot-file}

This is an interesting one since I see both the kernel and virtual hard drive get transferred to the target computer, but the kernel never starts.

The primary issue you have is this. The fog client runs as SYSTEM on the target computers. SYSTEM is an account with local administration rights, but doesn’t have any domain level rights (needed to access off system resources). If you need to access off system resources you will need to provide that connection information in the batch file, which from a security stance is not always the best choice.

As an example here is a batch file that will run an installer from a CIFS (SMB) share.

@echo off
net use w: \\server01\share01 /user:domain\user01 p@ssW0rd
w:
cd installers
setup.exe /qn!
net use w: /delete

As you can see in the made up batch file you would use the net use command and then supply the credentials needed to connect to an external resource. By doing that then SYSTEM can reach outside of the target computer.

With that said, its much easier some times to just create a snapin pack (think zip file) to install applications so the installer won’t have to reach outside of the target computer to get all of the needed files. I understand that creating a snapin pack is not realistic in all cases (think of AutoDesk design suite and how big the installer files are > 35GB). In that case having a shared install directory is the only/best solution.

@kiwijames Are these the storage node specs or is the FOG server running in virtual box? The storage statement has me a little confused too. How does win7 shared storage come into the picture? FOG is linux based. You have to do something abnormal to use CIFS for storage.

Can you explain a bit more about how/why you are using grub?

I don’t have my flash drive with me but if I remember correctly you can chainload from grub to tftp://<fog_server_ip>/ipxe.efi to chain from grub into FOS’s iPXE menu.

@imagingmaster21 Ok so the target computer is saying it can’t reach your image store. All of the kernel parameters look good.

So we need to do some checking.

1. Did you disable the firewall on your fog server, as well as selinux set to permissive as in the prerequisites to install FOG?
2. Is there some kind of firewall between the target computer and the FOG server?
3. Execute this command on your FOG server, showmount -e 127.0.0.1 and post the output here. (Hint if you connect to your fog server using putty you can copy and paste the text without needing to do screen shots).
4. Also post the output of this command from your FOG server ls -la /images

Are you spinning up a new server as in the a to b or adding a second fog server because you were asking about coping all images, not about server migration.

What is the IP address of the B server?

@edv-schuster From your picture it does appear that spanning tree is getting in your way.

As a test you can do 2 things, either one will tell us the same answer.

1. At the screen in the picture you gave, press the s key to get an iPXE shell. Wait 30 seconds and then key in dhcp net0 to see if it picks up an IP address. If not then key in dhcp net1, just in case the second network adapter is being used.
2. Place an unmanged (dumb) switch between the pxe booting computer and the building switch. This unmanaged switch typically will not support spanning tree. See if the computer boots properly into the FOG iPXE menu.

If either of the above tests point to spanning tree, then reinspect this switch to ensure that RSTP is enabled on all switch ports.

RSTP uses optimistic blocking (forward first then decide if there is another switch connected).
STP uses pessimistic blocking (block first then decide if there is another switch connected).

For what ever reason your switch might be blocking first then forwarding (at least based on the picture you provided).