@Jbob oh sorry, that was kind of vague.
It starts and all that but it doesn’t seem to do anything when I set the client to a task on the server.
This could be just me and I need to test more though.

@Jbob said:

@Arrowhead-IT what issues are you experiencing with including the client in the image? Its actually safer / more secure to install the client into the image rather than after imaging completes.

The client just plain doesn’t work when I include it in the image. Maybe I’m missing something, and I’m still doing testing. This is a new setup. At my old job I used to always include it in the image, but it had sporadic success rates too. I find that a fresh install always seems to work. Also, I love that I can install the client and it adds itself if the client isn’t already in fog. I figured that functionality meant that I should install it individually.

Although I imagine you figured this out by now, I also wanted to do this and found this post. I figured this is as good a spot as any to post it.
I’ve noticed that the service works better when it’s installed on each computer after imaging rather than trying to include it as part of an image.
So the idea of this script is to download the latest version and install it silently right after imaging, so I would suggest adding it as a snapin that you put on all clients. The only change you should need to make is changing the fogip variable to your fog server ip.

@ECHO off
set fogip=192.168.100.100
echo "downloading latest client..."
powershell -Command "wget http://%fogip%/fog.client/download.php?newclient -OutFile C:\FOGService.msi;"
echo "installing client..."
msiexec /i C:\FOGService.msi /quiet WEBADDRESS=%fogip%
echo "starting service..."
net start FOGService
echo "Done!"
exit

@kbramhall said:

@Arrowhead-IT As we’re troubleshooting I’m thinking it is definitely something with our internal networking configuration. The DNS servers we are using include google’s 4.2.2.2 and an internal DNS server.

Isn’t google dns server 8.8.8.8 and 8.8.4.4 ?
And if you’re using public dns’s maybe try opendns and see if that makes a difference
208.67.220.220 and 208.67.222.222

When you say internal network config, do you mean at the network infrastructure level or with the FOG server OS?
If you think its the local network, do you have any other cent OS boxes, and if so are they able to install the packages?

@kbramhall said:

@Tom-Elliott Hey Tom, we are not dead set to use 1.2.0 we just figured we’d download the latest from the FOG project website

Welcome to Fog, where if Tom were to release every stable release of the development branch it would be worse than Ccleaner updates.
1.2.0 was the last time something was labeled stable. I use the trunk in production and rarely have any issues, and you get new features all the time! And if there ever is a problem it’s fixed in a snap. Only issue with your configuration and the trunk is that you wouldn’t have internet, so you wouldn’t be able to install and updates. Which overall would be fine, I’m one of the crazies on the bit torrent sync to the source files because I can’t wait the 20 seconds it takes Tom to hit the commit button.

Anyway to install the trunk you’ll need git, svn, or btsync. Git is probably the easiest, svn is on sourceforge.net and they’ve gone down like 5 times in the last 3 months. There’s instructions on the wiki here https://wiki.fogproject.org/wiki/index.php/Upgrade_to_trunk

@kbramhall said:

@Tom-Elliott I attempted to go through the installer and saying no to DHCP and DNS but if failed to install tftp-server this time. Attached is the foginstall.log file.0_1450298881023_foginstall.log

I took a look at the install log and noticed this bit

../lib/redhat/functions.sh: line 1: n#: command not found

I’ve seen that before. You need to both make sure that you’re running as the root user and make sure you’re running the install script from the bin folder.
i.e. cd into where you downloaded/untarred the fog installer and then

cd bin
./installfog.sh

I figured out when making the automated update scripts that you can’t run it with the full path like
/home/fog/installFoder/bin/installfog.sh
because it use the trailing … to get to some included scripts. So you have to start the script from its happy home.

Also, what happens when you try to install the packages that failed manually?
I would try them one at a time. It looks like these ones…

yum install tftp-server
yum install xinetd
yum install vsftpd
yum install gcc
yum install gcc-c++
yum install lftp

And I just had another thought, are you sure the firewall is completely disabled? I just remembered a recent experience where a fresh install cent OS wouldn’t do internet things until I flushed the iptables.
Which if memory serves is

iptables -F
or
iptables -f

Hopefully something there helps

@Wayne-Workman said:

@Arrowhead-IT said:

Just so you know. The reason I chose not to use DHCP with FOG is because I had the mistaken assumption that FOG would automatically manage it in some way, or that the gui would have some control over it. It does not, you’ll be doing a bunch of manual configuration.

I guess now is as good a time as ever to say I’m working on an add-on that does just this.

@Wayne-Workman ooooh! I like that idea!

@kbramhall Considering that you only just set this up, maybe it’s worth trying a different distro. I’ve had better experiences with ubuntu server and fog then cent OS. But I am more familiar with debian so I might be biased.
But maybe it’s an issue with the OS install.
What happens when you run

yum update

?
If that doesn’t seem to connect to anything, then
What does your resolv.conf say?

cat /etc/resolv.conf

Might be a dns server thing. That’s sometimes the issue when you can’t seem to install a package

@george1421 said:

ok then lets confirm a few things.

You want FOG to be the dhcp server for your subnet?

Just so you know. The reason I chose not to use DHCP with FOG is because I had the mistaken assumption that FOG would automatically manage it in some way, or that the gui would have some control over it. It does not, you’ll be doing a bunch of manual configuration. It added some small configuration and did get it passing out ip addresses, which may be enough for you. But just be aware that you’ll likely have to do some more manual configuration of the DHCP. Not that that is all that hard nor undocumented on the internet, just wanted to make sure you’re aware.

@kbramhall
I’m not 100% sure if the package is the same in redhat as it is in debian, but in debian its
sudo apt-get isc-dhcp-server

so maybe try

yum install isc-dhcp-server

I’ve had a problem where it wasn’t installing the dhcp package on debian too on a test DHCP style install a couple weeks ago and installing the package manually fixed the problem for me. But then I ended up not using the DHCP on fog.

what dhcp package are you trying to install with yum?
And, since you’re new to the install, are you sure you want DHCP on the FOG server? Do you not already have a dhcp server on your network?

@Tom-Elliott said:

32 bit Init: https://mastacontrola.com/init_32.xz
64 bit Init: https://mastacontrola.com/init.xz

Just in case someone else is reading this and having the same problem and is new to fog and ipxe and what not…
To test out this fix do this…

cd /var/www/fog/service/ipxe
sudo mv init.xz init.bck
sudo mv init_32.xz init_32.bck
wget https://mastacontrola.com/init.xz --no-check-certificate && wget https://mastacontrola.com/init_32.xz --no-check-certificate
sudo chmod -R 775 ./

@Tom-Elliott is on top of this problem
Check out this forum post https://forums.fogproject.org/topic/6315/hp-z640-nvme-pci-e-drive/37
I’m posting my results of testing on the /dev/nvme0n1 type drives with his current inits found here

32 bit Init: https://mastacontrola.com/init_32.xz
64 bit Init: https://mastacontrola.com/init.xz

To use them do the following

cd /var/www/fog/service/ipxe
sudo mv init.xz init.bck
sudo mv init_32.xz init_32.bck
wget https://mastacontrola.com/init.xz --no-check-certificate && wget https://mastacontrola.com/init_32.xz --no-check-certificate
sudo chmod -R 775 ./

Granted by the time dell gets you a replacement it might already be tested enough and the fix will just be in the latest trunk. But that’s what to do in case anyone else is looking.

@Tom-Elliott
Ok I will test every which way just for funzies and report back by editing this post

I won’t test the multiple disk images because I only have one disk in there right now. But I can add in a drive that would get recognized as /dev/sda and make sure it can image that combination if you want.

Download
Single Partition/Resizable - Working!
Multiple Partition/Single Disk - Working!
Multiple Partition/Multiple Disk - Working!

Upload
Single Partition/Resizable - Working! - a few error messages that don’t stop anything
Multiple Partition/Single Disk - Working!
Multiple Partition/Multiple Disk - Working!

@Tom-Elliott
Sorry that it took me so long, I kept getting distracted by some other stuff.
It’s Alive!
It Works!
Amazing!
You’re the best!

• I updated to the latest btsync (5750) (which seems to be newer than the latest git of 5748)
• Downloaded your inits from mastacontrola
• I also took out the host primary disk specification in the gui for gits and shiggles
• I tested download with the already existing single partition resizable base image.
• It worked perfectly!

I did a debug deploy so I can also confirm that all the variables are now correct…

Shall I test a resizable and a not so unresizable upload?

I’m guessing that you did a uefi image since it wants to use windows boot manager, which I believe is window’s uefi method.
Was secure boot enabled on the computer that you uploaded from, could that cause a problem? I haven’t actually done any successful uefi imaging and haven’t really tried that hard because I’m personally yet to see why it’s worth it when mbr/bios/legacy works so well and I don’t have any 3+TB drives in my windows workstations at work.

I have had success imaging windows 10 with legacy mode, just for a reference, I’m on the latest trunk, but I would bet 1.2 works with it.
Granted I don’t use sysprep, I find that it likes to kill off my default profile. So I just manually uninstall devices in the device manager before imaging. It’s trickier, and involves more work, but it’s done sooner than it takes sysprep to run. But I’m getting off topic.

Have you tried repairing the install with a win 10 disk. Auto repair. What ever the gpt/uefi equivalent of

bootrec /rebuildbcd
bootrec /fixboot
bootrec /fixmbr

is. (Fun fact shift+F10 gets you straight to a command line on a windows install disk instead of waiting to click through the menus)
Or try the windows automatic repair that works like 10% of the time.
You can also use diskpart to make sure the volume/disk/partition that got imaged is active.

disk part
select disk 0
list partition
select partition 1
detail partition

something like that, and then you can set it to active. That was a problem I used to have when I used the altiris imaging system, I’ve never had the problem with FOG.
Granted you shouldn’t have to repair the install after imaging, especially if you run sysprep all proper, but if you can make it work then that might narrow down some possibilities.

Can you disable secure boot in the bios settings on the computer not booting?
Can you make sure secure boot was disabled on the computer you uploaded from, or if it was disabled try imaging with it enabled?
I’m just guessing on the secure boot thing, might not even be a thing, but it seems like it could be relevant.

I hate to suggest starting over on an image, but I would recommend making sure that you’re using bios/legacy/mbr boot mode. Unfortunately there is no way (that I have ever found anyway) to switch a computer between the two without reinstalling the OS because they use different partitioning types MBR and GPT. Also, when installing the OS in mbr mode, the easiest way I’ve found to ensure it’s a single disk (This hasn’t worked for uefi/gpt mode when I’ve tried) is to hit shift+f10 at the partition screen then do the following in the command line to clear the partitions and make only 1 full size partition

diskpart
list disk
# select your disk, usually disk 0 unless you have more
select disk 0
clean
create partition primary
exit
exit

Hit refresh on the partition selection screen and viola. I got that from a forum post on here or related to fog somewhere many moons ago and have been using it forever.

I seem to be rambling again.
To sum up

• Figure out if you’re in uefi or legacy mode
• If you are in uefi mode, play with secure boot and try a repair disk
• If you can, remake the image in bios/legacy/mbr mode
• Hopefully something I said is helpful

@Wayne-Workman That makes me feel special Thanks

@Tom-Elliott I’m giving it a go now!

Have you tried a proxy dhcp using dnsmasq like they’re talking about?

I do realize that you are using a redhat not a debian, so unless you switch to a debian flavor of linux like ubuntu server or what have you, the commands would be slightly different, but I’m pretty sure that there should be equivalents for this all in redhat if you don’t want to switch distros again, and I wouldn’t blame you at all.

I only know how to set it up on debian based linux with apt-get, but I imagine the yum RHEL package is probably the same name.
From what I recall to set it up…

sudo apt-get update
sudo apt-get install dnsmasq
sudo nano /etc/dnsmasq.d/ltsp.conf

In the ltsp.conf file put this…

port=0
log-dhcp
tftp-root=/tftpboot
dhcp-boot=undionly.kpxe,fog-server,10.1.8.1
dhcp-no-override
pxe-service=X86PC, "Boot from network", undionly
dhcp-range=10.1.8.1,proxy

You may need to make symlinks for dnsmasq to see the undionly file, and I’ve had some luck putting copies of the bootfile in the ipxe folder too, that very well may do nothing, but I think it fixed something once so I just kinda still do it to be safe since it doesn’t hurt anything.

sudo cp /tftpboot/undionly.kpxe /tftpboot/undionly.0
sudo cp /tftpboot/undionly.kpxe /var/www/fog/service/ipxe/undionly.kpxe
sudo cp /tftpboot/undionly.kpxe /var/www/fog/service/ipxe/undionly.0

Then restart the dnsmasq service (also do this whenever you make changes to this file)

sudo service dnsmasq restart

You can play with the dhcp-range value a bunch to try to get it to work. This is the method I used when I didn’t have access to the DHCP server.
You can try setting the range to just your FOG server’s ip. You can set it to the router gateway, you can set it to the dhcp server too. And you can try interchanging I’ve had some

In the event that enabling dnsmasq kills your resolvconf dns configuration (which it often does…) the easiest solution is to add your dns servers in /etc/resolvconf/resolv.conf.d/base
You would also want to edit /etc/resolvconf/interface-order to have your main network interface at the top. It defaults to local host first and makes it so 127.0.0.1 becomes your dns server and breaks the internet. At least it’s done that every time I’ve installed it on ubunutu.
so like this…

sudo nano /etc/resolvconf/resolv.conf.d/base

Then lets say you use google’s dns servers and google.com was your domain for a simple syntax example, you would do this…

nameserver 8.8.8.8
nameserver 8.8.4.4
search google.com

Then restart the resolvconf service

sudo service resolvconf restart

If that doesn’t take care of it and it is a problem, check out this forum post for more info
https://forums.fogproject.org/topic/2799/dnsmasq-kills-dns-lookup-on-fog-server/6

A basic rundown of my understanding of dnsmasq is this

• Computer tries to boot to pxe and starts looking for the dhcp server
• Your fog server pretends to be a dhcp server and beats the real dhcp server to it (If your fog server is on a slower switch than your dhcp server i.e a 100 Mbps port and the infrastructure is on a 10 Gbps port, then this might not work, I had that problem once)
• The Dnsmasq service gives the computer the bootfile and uses the real dhcp server to get an ip address
• You boot into pxe and do a happy dance

More information on dnsmasq and fog here https://wiki.fogproject.org/wiki/index.php/Using_FOG_with_an_unmodifiable_DHCP_server/Using_FOG_with_no_DHCP_server
and here https://wiki.fogproject.org/wiki/index.php/ProxyDHCP/dnsmasq-_DRAFT

Another option, albeit an extremely advanced and difficult option, is to set up fog as a tertiary dhcp server. You would have to look into configuring a linux dhcp server, not something I have personally done at that in depth of a level. Because you would have to figure out how to make it act as a separate vlan or something like that, or maybe just have a separate network for imaging rather than having it on your whole organization. Something like a switch and a workbench in your office. But those are just a couple fall back ideas.

Also a side note, @george1421 and @Wayne-Workman Wayne is right here. I’ve actually gotten dnsmasq to work on multiple subnets and buildings. Granted it was a complicated university environment, and I had to do some tricksy stuff. But don’t underestimate the power of dnsmasq.

@Tom-Elliott The first test I did was with a multi-disk image that I made from a vm and it went back to the behavior it did before where it acted like it was working but never gets into partclone.
I’ll do a debug session for that one since it happened to fast to snap a picture.