I did some further testing.
With BIOS Mac-address pass-through set to ‘disabled’ - ipxe.efi works fine.
However with BIOS Mac-address pass-through set to ‘system mac’ - ipxe.efi fails as per the original posters screen shot of ‘no configuration methods’. However using snponly.efi did work in this mode
I had the same battle with these 450 G8’s, and new HP Elitedesk 600 G6 micro desktops. But I now have it working!
My real problem was rEFInd hanging after the PXE boot menu. I updated /var/www/fog/service/ipxe/refind.efi on the primary fog server, with the copy from the latest FOG 1.5.9 zip. (The refind.efi copies on any storage nodes isn’t used?)
I’m PXE booting with ipxe.efi also from FOG 1.5.9
Then in the HP BIOS, the HP SureProtect stuff was all disabled too
Unrelated but the bzImage kernel also needs to be the latest build so the NIC is detected on it. The new February 2021 build solved that one.
EDIT: I’m sure I’ve seen your original error message too about no configuration methods. I can’t remember if this was solved via BIOS update - or if it was more switch-port related.
Hi Guys
Two quick updates on this:
This ipxe7156.ipxe also seems to be working for some Dell machines we have.
To make Hyper-V Generation 2 VM’s to PXE boot UEFI correctly; I had to add a new PXEclient mapping of PXEClient:Arch:00007:UNDI:003000 for the ipxe7156.ipxe boot file. (Secure boot should also be disabled!)
(Tested with FOG 1.5.2)
Revised Powershell Code:
import-module dhcpserver
$DHCPserv = "DHCP-SERVER-HOSTNAME"
Add-DhcpServerv4Class -Name "PXEClient (UEFI x64) 2" -Type Vendor -data 0x505845436c69656e743a417263683a3030303032 -Description "UEFI Architecture Type 2" -ComputerName $DHCPserv
Add-DhcpServerv4Class -Name "PXEClient (UEFI x64) 6" -Type Vendor -data 0x505845436c69656e743a417263683a3030303036 -Description "UEFI Architecture Type 6" -ComputerName $DHCPserv
Add-DhcpServerv4Class -Name "PXEClient (UEFI x64) 9" -Type Vendor -data 0x505845436c69656e743a417263683a3030303039 -Description "UEFI Architecture Type 9" -ComputerName $DHCPserv
Add-DhcpServerv4Class -Name "PXEClient (UEFI x64) 7" -Type Vendor -data 0x505845436c69656e743a417263683a3030303037 -Description "UEFI Architecture Type 7" -ComputerName $DHCPserv
Add-DhcpServerv4Class -Name "PXEClient (UEFI x64) 7 UNDI 3016" -Type Vendor -data 0x505845436c69656e743a417263683a30303030373a554e44493a303033303136 -Description "UEFI Architecture Type 7 UNDI 3016 (Surface Pro 4)" -ComputerName $DHCPserv
Add-DhcpServerv4Class -Name "PXEClient (UEFI x64) 7 UNDI 3000" -Type Vendor -data 0x505845436c69656e743a417263683a30303030373a554e44493a303033303000 -Description "UEFI Architecture Type 7 UNDI 3000 (Hyper-V)" -ComputerName $DHCPserv
Add-DhcpServerv4Policy -Name "UEFI-IPXE" -Description "UEFI IPXE" -ScopeId 0.0.0.0 -Condition OR -VendorClass EQ, "PXEClient (UEFI x64) 7*","PXEClient (UEFI x64) 9*" -ComputerName $DHCPserv
Add-DhcpServerv4Policy -Name "UEFI-IPXE32" -Description "UEFI IPXE32" -ScopeId 0.0.0.0 -Condition OR -VendorClass EQ, "PXEClient (UEFI x64) 2*","PXEClient (UEFI x64) 6*" -ComputerName $DHCPserv
Add-DhcpServerv4Policy -Name "UEFI-IPXE7156" -Description "UEFI IPXE7156" -ScopeId 0.0.0.0 -Condition OR -VendorClass EQ, "PXEClient (UEFI x64) 7 UNDI 3016*","PXEClient (UEFI x64) 7 UNDI 3000*" -ComputerName $DHCPserv
Set-DhcpServerv4OptionValue -PolicyName "UEFI-IPXE" -OptionId 67 -Value "ipxe.efi" -ComputerName $DHCPserv
Set-DhcpServerv4OptionValue -PolicyName "UEFI-IPXE32" -OptionId 67 -Value "i386-efi/ipxe.efi" -ComputerName $DHCPserv
Set-DhcpServerv4OptionValue -PolicyName "UEFI-IPXE7156" -OptionId 67 -Value "ipxe7156.efi" -ComputerName $DHCPserv
Set-DhcpServerv4OptionValue -OptionId 67 -Value "undionly.kpxe" -ComputerName $DHCPserv
Set-DhcpServerv4OptionValue -OptionId 66 -Value "10.20.30.40" -ComputerName $DHCPserv
I’ve had great success with a “universal” image so far on legacy machines. But I agree; now that I’ve done further research, any hybrid image with UEFI is not looking good. (Although converting the SOE master VHD image to a WIM & transposing that to a USB has worked relatively well despite not being fully automated by FOG’s management.)
As it turns out MBR2GPT.EXE via WinPE failed to find the OS partition - though it was there so FOG did it’s job.
Either way I give up - wasting too much time on this - 2 x images it is 
gdisk appears to have worked on the Dell - no errors - but the ‘BIOS’ setup on the dell doesn’t see any UEFI boot loader.
I might be better off just building a WinPE with MBR2GPT.EXE in it… or just resign to the fact of maintaining dual images.
Was hoping to have a universal image that could be deployed to either a legacy or UEFI machine. The bulk of our fleet is legacy/MBR booting; but we’ve got new Dell 7390’s that refuse to boot legacy off the internal disk. Also have some Surfaces which we’ve been doing manually via USB WIM’s (converted from the SOE master HyperV VHD)
So now trying to avoid having to maintain another image in UEFI format for these. Of course I could just switch the existing fleet to UEFI boot, but that’s got it’s own logistical and possible technical/compatibility issues to deal with.
Have used MBR2GPT.EXE on a legacy-booted machine, restarting & changing BIOS to UEFI mode then resulted in a working UEFI machine; so I was hoping there might be some sort of linux based conversion possible!!
The post-download script looks perfect - I could easily script from that!
However testing gdisk (on a Surface Pro 3 which was just deployed with a MBR based image), resulted in errors about overlapping partitions - which were detected incorrectly anyway. About to see if the Dell does the same thing.
The Wiki’s info on UEFI setup is great, but fiddly to setup on lots of Windows Servers, so I’ve converted it to PowerShell.
Suggestions/corrections welcome - the architecture selection is somewhat hard to test with limited devices.
Hope this helps others!
import-module dhcpserver
$DHCPserv = "DHCP-SERVER-HOSTNAME"
Add-DhcpServerv4Class -Name "PXEClient (UEFI x64) 2" -Type Vendor -data 0x505845436c69656e743a417263683a3030303032 -Description "UEFI Architecture Type 2" -ComputerName $DHCPserv
Add-DhcpServerv4Class -Name "PXEClient (UEFI x64) 6" -Type Vendor -data 0x505845436c69656e743a417263683a3030303036 -Description "UEFI Architecture Type 6" -ComputerName $DHCPserv
Add-DhcpServerv4Class -Name "PXEClient (UEFI x64) 9" -Type Vendor -data 0x505845436c69656e743a417263683a3030303039 -Description "UEFI Architecture Type 9" -ComputerName $DHCPserv
Add-DhcpServerv4Class -Name "PXEClient (UEFI x64) 7" -Type Vendor -data 0x505845436c69656e743a417263683a3030303037 -Description "UEFI Architecture Type 7" -ComputerName $DHCPserv
Add-DhcpServerv4Class -Name "PXEClient (UEFI x64) 7 UNDI 3016" -Type Vendor -data 0x505845436c69656e743a417263683a30303030373a554e44493a303033303136 -Description "UEFI Architecture Type 7 UNDI 3016 (Surface Pro 4)" -ComputerName $DHCPserv
Add-DhcpServerv4Policy -Name "UEFI-IPXE" -Description "UEFI IPXE" -ScopeId 0.0.0.0 -Condition OR -VendorClass EQ, "PXEClient (UEFI x64) 7*","PXEClient (UEFI x64) 9*" -ComputerName $DHCPserv
Add-DhcpServerv4Policy -Name "UEFI-IPXE32" -Description "UEFI IPXE32" -ScopeId 0.0.0.0 -Condition OR -VendorClass EQ, "PXEClient (UEFI x64) 2*","PXEClient (UEFI x64) 6*" -ComputerName $DHCPserv
Add-DhcpServerv4Policy -Name "UEFI-IPXE7156" -Description "UEFI IPXE7156" -ScopeId 0.0.0.0 -Condition OR -VendorClass EQ, "PXEClient (UEFI x64) 7 UNDI 3016*" -ComputerName $DHCPserv
Set-DhcpServerv4OptionValue -PolicyName "UEFI-IPXE" -OptionId 67 -Value "ipxe.efi" -ComputerName $DHCPserv
Set-DhcpServerv4OptionValue -PolicyName "UEFI-IPXE32" -OptionId 67 -Value "i386-efi/ipxe.efi" -ComputerName $DHCPserv
Set-DhcpServerv4OptionValue -PolicyName "UEFI-IPXE7156" -OptionId 67 -Value "ipxe7156.efi" -ComputerName $DHCPserv
Set-DhcpServerv4OptionValue -OptionId 67 -Value "undionly.kpxe" -ComputerName $DHCPserv
Set-DhcpServerv4OptionValue -OptionId 66 -Value "10.20.30.40" -ComputerName $DHCPserv
Also, for those trying to get UEFI working correctly, I had to also do the following:
Can we get the ability to automatically convert a MBR based image to GPT after a deployment?
Looks like it’s possible via gdisk without data loss. If this could be scripted after a clone job, that would be a miracle cure for me.
More info here: http://slavisa-jovanovic.com/linux/2015/02/19/mbr-to-gpt.html
We use http://www.sivann.gr/software/itdb/ with a few minor UI/language mods. Works a treat!
We’d used Spiceworks in the past, but it’s dynamic/scanning nature tended to mess things up. This sometimes added great confusion as to what the actual truth was.
With a completely manual database, at least we know any changes would’ve been done by a human and not some script editing our records. FOG’s inventory is still good for cross-checking the asset db records though!
Big FOG user here with many locations across Australia - some 3000+km with storage nodes/sites inbetween
Having audit/history logging for hostname and Active Directory changes to the FOG database would be very handy. Had a couple of instances where I’ve needed to identify which tech has modified a particular host. eg: Hostname and/or AD OU has ‘magically’ changed but I can’t pin it down to which tech did it.
Either way, keep up the good work!
No SSD drives, they are standard Optiplex 3010’s. Will give the delayed started trick a go.
Imaged a lab of 25, and I had 5 machines fail to rename. Kicking them off again they then decide to work fine. Seems to be happening randomly - i am having to double check that everything has actually finished / domain joined correctly still.
Logging in to the failed machines locally, they still have the sysprep generated PC name.
Changing the FOG rename early setting hasn’t made any difference.
Hello
Can I wish for a task button where by a computer can be forced to rename itself & join the domain again?
Much like the ‘Configuration Task’ (?) back in Ghost Solution Suite.
This would be handy for any machines that decide not to rename/join domain for whatever reason. Be easier to kick this off rather than an entire reimage.
Odd thing too is the problem lab was still booting Win7 fine despite gparted showing a blank partition table too! Back to reimaging now
Ah man… :eek:
Just ran gparted over it and it said there was no partition table. Created it with gParted and then rebooted, and it’s imaged up fine then. Which also explains why a brand new (replacement) drive didn’t work.
I’d just assumed that FOG would restore all that from the image too. (mult-partition image).
Also if I tell it to do a Memtest task that works fine. Just seems to totally skip the imaging step.
Hi Tom,
I’ve tried that and it’s no different.
Looking at the apache log on the primary server, I’m seeing this when I PXE boot a problematic machine, then select the ‘quick image’ task.
10.1.1.110 - - [28/Oct/2014:12:59:59 +1000] “POST /fog/service/ipxe/boot.php HTTP/1.1” 200 609 “-” “iPXE/1.0.0+ (3a02)”
10.1.1.110 - - [28/Oct/2014:12:59:59 +1000] “POST /fog/service/ipxe/boot.php HTTP/1.1” 200 948 “-” “iPXE/1.0.0+ (3a02)”
10.1.1.110 - - [28/Oct/2014:13:00:08 +1000] “POST /fog/service/inventory.php HTTP/1.1” 200 299 “-” “Wget”
10.1.1.110 - - [28/Oct/2014:13:00:08 +1000] “GET /fog/service/Pre_Stage1.php?mac=78:45:c4:2f:61:09&type=down HTTP/1.1” 200 300 “-” “Wget”
10.1.1.110 - - [28/Oct/2014:13:00:08 +1000] “GET /fog/service/Post_Stage3.php?mac=78:45:c4:2f:61:09&type=down HTTP/1.1” 200 297 “-” “Wget”
Update:
ImagingLog db table shows an entry for that machine ID with identical start/finish times too, along with the TaskLog table showing State 3 & 4 entries with identical times too.
Hi All,
Been using FOG in a multi-site setup for the past year quite successfully. Some 800+ jobs have been kicked off.
The main FOG server is located in our data centre and we have local storage nodes each at each site.
Recently during a unicast clone session at one site, the VPN link dropped out whilst imaging was taking place. When these machines finished imaging (from the local storage node) they then bombed out with an error message because the primary FOG server couldn’t be contacted. (Sorry I don’t recall what this message was!)
Once the VPN was back up, the jobs were cancelled. That actual deployment still worked OK, but the couple of machines we have since attempted to reimage are now not working. The machines still PXE boot in to the FOG OS, then instead of partclone executing, it appears to skip straight over it and then reboot. FOG then thinks the job is completed successfully.
My guess is some temporary file/flag is set somewhere still? Any ideas of where I should start looking to clean this up?
EDIT: Running Fog 1.2.0 too
Also, documentation wise - unless I was blind/missed them elsewhere - the following Migration steps are missing from this page [url]http://www.fogproject.org/wiki/index.php/Upgrade_to_1.0.0[/url] /
Hopefully that will help others - it has been well worth the time to upgrade here
Hi
Fog 1.1.2 is looking like an awesome improvement over 0.32 so far - which in itself was an awesome upgrade over Ghost Solution Suite 
Not sure if this has already been reported, however FYI…
A) Renaming computers in the Host Management page usually doesn’t seem to save. Although other settings on that host/general page do save.
B) I installed the Location plugin on 1.1.2 (previously not installed when I upgraded from 0.32) and whilst the Location page icons appears, none of the “New Search”, “List All Locations”, “Create New Location” pages display anything. (I’ve just edited the database location table manually with the required details and this does appear to actually work, so I can assign hosts to locations and then they TFTP boot the bzImage from the correct server then :))
C) I don’t think Multicast is working at all - partimage just sits waiting. Not too worried as we mostly unicast anyway. (FYI on 0.32 Multicast was super slow/useless on 100Mbit workstations, but perfect on any Gigabit machines. Yet Ghost’s Multicast worked fine on the same machines. HP Procurve switches with IGMP filtering configured, etc.)