@george1421 Great you’ve been helping out so much here and figured it all out.

@mstumpo Marking this solved as I think the initial question(s) is/are solved. Please feel free to open new topics as you run into new issues. Good to keep topics clean so other people can find this and profit from the answers as well (details not lost in a 5 page topic).

First let me say that FOG will not work as you have it defined. FOG will only image using a single IP address on the FOG server. Your FOG server can have more than one network interface but will only image using a single IP address.

I know I’m mixing what I’m saying. FOG does support network bonding (using two or more physical network interfaces to add additional lanes of bandwidth). But those bonded interfaces MUST have a single IP address assigned. And typically the bonded interfaces go to a single network switch.

Your drawing appears you want a multi-homed FOG server (multi-homed means multiple network interfaces each with their own IP address). That is not supported for FOG imaging.

@jtyler Have you had a look into the mentioned error_logs? Let us know what you get.

I managed to make a simple symlink on the /usr/lib/systemd/system/ folder
php-fpm.service -> /usr/lib/systemd/system/rh-php56-php-fpm.service and this seemed to clear the problem.

thank you

@george1421 It’s now working, as i’ve installed so many times this time around I forgot to open the web panel and import the configuration 😉

thanks for your help

@unltdlove2u How did you create that ubuntu-16.04.img in the first place? Sounds to me as if this is just your KVM image file, right? Well with FOG the steps to create an image is: register the machine (as well KVM should work if all network and DHCP is setup properly), schedule a capture/upload task and let the client upload its disk contents.

What you are asking for is a way to convert your KVM disk image into a FOG image. Might be possible but it’s not supported. Search the forum for convert vmdk and such VM disks. I think some people have done it but it’s probably a lot of work and I am not sure it’s worth it.

For dm-crypt/LUKS you need to let FOG do a sector by sector copy (also called dd) as it can’t read the encrypted filesystem stuff. The FOG image will be very big as it contains unused sectors as well.

I’m having the same trouble with 18.04. Was curious if you found a work around. Booting the ISO itself works fine, but PXE booting doesn’t work. My guess is that the network, obviously, has to start sooner in order to pull over the files via NFS, and this is somehow mucking up the traditional order of operations.

@hvaransky I also recommend you open a new thread unless your fog server and configurations are exactly the same as the OP’s I don’t want to confuse posted information because we are talking about different setups.

@kte Read through this: https://wiki.fogproject.org/wiki/index.php?title=Troubleshoot_FTP

Use the built in kernel upgrade tool in FOG Settings -> FOG Kernel to install the down level kernels.

In the past I’ve had issues with the FOG server being able to connect to the internet through our departments firewall/proxy systems. I have managed to get it working this time tho thankfully and have downgraded the kernel!

I have also edited the php values and changed the client check in settings to 300. This has helped the CPU load on the server although it still seems a little high than what we are used too. At least it is a lot more responsive now on the web interface!

We are now having a GPT issue deploying an updated image that was captured after the upgrade (pre-upgrade images seem to deploy fine). I have started a new forum post on this though.

there isn’t anything there yet:
anthony@fogserver2:/opt$ ls
1.5.4.tar.gz fogproject-1.5.4

@walkerit To expound a bit on @Quazz 's comment, I ran into this yesterday myself. There’s two different config locations that I found needed to be updated. One of them I’m led to believe will only come into play if you update the server, but the other (export? https://wiki.fogproject.org/wiki/index.php?title=Change_NFS_location) is what was causing my permission issues.

Hope that clears it up when you do change your storage location.

Philip

An update, solved.

There were two issues.

The timezone issue was resolved by setting the Centos system to to UTC and Fog’s ‘TZ INFO’ setting to UTC. Don’t know why setting to ‘UTC’ worked and not with the same settings using ‘EDT’ <shrug>. Once we catch up on our imaging I’ll try to revert back to our timezone and see if this is merely an annoying warning or if it still causes a fatal error.

The second issue was the user ‘fog’ passwords were not consistent, system vs. Fog account. So when Fog completed the imaging and during the ‘Updating Database’ routine the ftp login failed and threw ‘Error returned: Type 2’.

0_1528461696724_IMG_0807.JPG

We are now able to image Window 10 clients for the first time! Thanks everyone for your help!

Scott

@timo You can restart apache2, the effects won’t last long though. When you can get into the web gui, you should go to Storage Management -> your storage node -> Max Clients and set this to something like 1 or 2. This will limit how many snapins run at once - and will allow your fog server to operate normally while the snapins run.

@stylus3530 In fact, you can deploy a 1.2.0 image and re-capture with 1.5.2. All you need to do is not allow the image to boot. Once deployed, immediately re-capture and all should work fine.

@kleber I’d recommend removing the background image via the GUI under Settings | iPXE General Config. I just removed the background image name from that field entirely. The default background prevented certain machines from being able to boot for me as UEFI. Figure its a quick enough thing to check.

Are there other machines that do boot ok with your setup or are you not able to boot any machine via iPXE?

Do the machines you have issues on have any BIOS settings specific to PXE booting (enabling it, authentication, etc)? If you you may want to check the configuration there.

Have you tried UEFI boot instead of BIOS boot, I have found sometimes the error is more descriptive in one vs the other (or even that one way has no error and the other does). I have found it especially helpful to have a camera/phone camera with high speed video recording, ie 120/240 fps and record the screen after I make my selection to catch any brief messages on screen.

@tom-elliott Cool, that worked perfectly 🙂

@stylus3530 yes the client does have renaming for Linux, Mac OS, and Windows. Windows has the benefit, however, of the host name early working from the imaging context.

@veresh The commit says the packages were added to enable compilation of iPXE binaries. As far as I understand the compiled iPXE binaries are (still?) pulled from the repository anyway, so I guess it’s to allow for updating (or modifying) the binaries manually without updating FOG

@george1421

I found my mistake

The IP was wrong the old VM snapshot had NAT instead of bridged network interface enabled

Thank you very much for the fast support.

alt text