i’ve noticed that when creating an image from the ground up, renaming the computer BEFORE capturing an image could cause problems. for some reason the netbios name of the original name that was created by the windows setup process will stick even though you renamed the pc. for example. i imaged a drive from our sccm server with a fresh windows 10 image, it had that minit-xxxx default name (cant remember exactly what it was but similar to that), so i changed the name to LAB-TEST then captured and deployed the image using acronis, renamed the pc to be unique to that workstation, then when the internet was connected, it pulled a major windows update (1607>1803). some of the workstations were automatically renamed back to the original image name that i set LAB-TEST. This caused trust relationship issues when trying to join to the domain. i found this entry in the registry HKLM\SYSTEM\CurrentControlSet\Control\ComputerName\Active ComputerName and ComputerName. The string that reads ComputerName in each of those keys had LAB-TEST. So i disjoined from our domain, and changed both entries to read what the unique name i had set was and was able to join the domain. so the lesson here was to not touch the computer name until AFTER you install windows, capture the image, and then deploy the image. hope this helps.

oh ok. i probably missed that detail on the ordering of ethernet assignments. maybe one day i can redo the server again and make it right.

but the UI is in fact accessible on both interfaces.

forgive me but i just want to clarify that the entire sql database consists of all of those exportable items found in the web ui?

My FOG server has two nics. eno1 and eno2. eno1 is bound to my standard workplace LAN and eno2 is set for DHCP and tftp for imaging and that network is physically isolated. The WebUI is accessible via both interfaces. I found a bug where if i log into the UI from the eno1 network, the storage node says its offline node unavailable, and if i click on it, it takes me to another page that says “Invalid Server Information!” The storage node is configured with the ip that is bound to eno2. i have a workstation set up for management on the same network that when accessing the web ui, the storage node shows up properly with space used/available. Now this bug isnt exactly persistent because sometimes it shows up properly when accessing via the eno1 network, and sometimes it doesn’t so its hard to reproduce and troubleshoot. hope this makes sense.

So i ran into an issue recently where i had to reformat my server, reinstall FOG, and import the following from backups that i had:
Users, Hosts, Groups, Images, and FOG Settings. I manually backed up /images to a different workstation with plenty of space. I used the export function to create these backups. My coworker had registered a new host and captured an image after i had created those backups. So he had to re-do that work. So is there a way to automatically backup those items on schedule? Not sure if using mysql will export all of those items. If it does then i can just create a cronjob to complete that task. I will run it every weekend when no one will be capturing or deploying images.

Right my freenas server is hosting on http not https via webdav. it is being hosted on an isolated network so i dont see the need to secure it, plus the webdav is set to read only. so FOG is set to ipxe over https by default? how can trust be established at the BIOS or UEFI level?

the only time ive seen permission issues with ipxe is when the tftproot folder isnt public with read and execute, but that was on my freenas server using dnsmasq for dhcp and tftp. maybe check the permissions on /tftproot and /var/www/fog/service/ipxe.

how is FOG able to set the hostname in Windows immediately after imaging though? I’ve noticed that after imaging, the computer name in Windows is set to whatever i set the hostname to during the Full Registration process.

Great thanks! should i use 1.5.7 or the latest git version 1.5.7.2?

i did read your warning loud and clear, but like i said the Web UI was down and i didnt want to leave the server down all day through the weekend and into this morning, so i rolled the dice. The process of reinstalling everything wasn’t too bad. hey now i have backups of everything and i learned a lesson.
Ok so if i just say “No” to the first part of the installation where it asks if i want to reinstall apache and php, everything else should go through without issue?