If you don’t have overlapping subnets between the two networks it may be easier to setup routing so that both networks can see and use the fog server.

I don’t know if there is a way to make Fog work the way you are proposing using another interface.

I user dual monitors, but I haven’t uploaded or deployed an image that was from a dual monitor setup. What issues are you running into?

Have you tried the InstallRite forums/support site?

To image a fog server, you’ll probably need something that supports ext4. Look into clonezilla, ghost, or acronis. I think you can always DD to a spare hard disk also

I use a pretty beefed up Fog server that has 6 disks in RAID 5 and dual gigabit nics bonded. I can unicast about 20 machines at a time. I am planning on adding a few spare desktops to my Fog setup as additional storage nodes to see how many more I can unicast at a time, but disk throughput in the bottleneck on the single-disk servers. More than 2 or 3 clients and they all slow down really bad.

I think you can set it up at your office with the IP address it will have at the other network. If you configure DHCP services, make sure you set it up for the other network also. The Fog server will not function at the office, i.e. you can’t really test it by imaging/registering computers, but you can once you get to the new network.

I would probably choose to set it up in the current network, verify it’s working, then move it to the new network by updating the IP address via the wiki article linked in Roberts response. Deploy the server and troubleshoot from there if needed.

Go to your Fog web ui, Other Information, Fog Settings, and find the TFTP Server section. YOu’ll need the FOG_TFTP_FTP_USERNAME and FOG_TFTP_FTP_PASSWORD values.

Go to Fog web ui, storage management, All Storage Nodes, and click on defaultMember. Make sure your management username matches the FOG_TFTP_FTP_USERNAME from the other screen. Same with Management Password and FOG_TFTP_FTP_PASSWORD.

Update, go back to home screen, see if you can connect to the defaultMember.

If you still have issues after this, we’ll continue, but try this to get rid of the errors about connecting to the storage node.

Have you checked any other clients for the ability to PXE boot or are you stuck with only GX520’s?

Have you already modified your Fog 0.32 on Ubuntu 10.04 to support chainloading?

Looking back up at the previous posts, the error “SubProcess -> mirror: Login failed: 503 Login with USER first” means you probably don’t have the Management username and password specified in your storage node definition.

If you don’t have the correct Management username and password in your storage node definition, you should see an error “SubProcess -> mirror: Login failed: 530 login incorrect”

Ok. This is what I’ve found in my Ubuntu 10.04 LTS Server running Fog 0.32.

1. Update MySQL to allow connections from remote hosts. On the main Fog server, edit /etc/mysql/my.cnf changing the line: bind-address = 127.0.0.1 to bind-address = [ip of server]
2. On the storage node, make sure the file: /opt/fog/service/etc/config.php has the fogstorage username and password to match the values from your main server. (Fog Web UI, Other Settings, Fog Settings, FOG Storage Nodes username and password).
3. On the storage node, open up the /opt/fog/.fogsettings and get the password from the line: password=“somepassword”;
4. On the main server, fog Web UI, define a new storage node (or update the one that is not working). Put in all the information like name, ip address, max clients, DO NOT CHECK “is master node” - see warning near box, select the storage group (default usually), image location to “/images/”, check “Is Enabled”, and put in the Management username as “fog” and management password as the value we got from the /opt/fog/.fogsettings file.
5. On the main fog server, run: sudo service FOGImageReplicator restart
6. On the mian fog server, run: tail -f /opt/fog/log/fogreplicator.log
7. verify your images start syncing.

I checked [URL=‘http://freeghost.svn.sourceforge.net/viewvc/freeghost/trunk/’]trunk [/URL] [URL=‘http://freeghost.svn.sourceforge.net/viewvc/freeghost/trunk/lib/’] > lib [/URL] [URL=‘http://freeghost.svn.sourceforge.net/viewvc/freeghost/trunk/lib/ubuntu/’] > ubuntu [/URL] > functions.sh (861). The installPackages() function still only echos out a note about updating the config file under ${webdirdest}/commons/config.php and mentions nothing about /opt/fog/service/etc/config.php.

iirc, both locations must be updated if you use a non-empty root password in mysql. The first for the web ui, the second for the FOGImageReplicator service.

There is a difference between the /etc/dnsmasq.conf and /etc/dnsmasq.d/ltsp.conf

Fog polls the computer for DMI information, including the Asset Tag info from the BIOS during registration, but after the host has been named. You probably just need to update the fog scripts inside the init.gz (/tftpboot/fog/images/init.gz) to update the database with the asset tag as the host name at the end of registration and before the machine is imaged.

This sounds like a possible plug-in/add-on idea.

I believe this has to do with the A50M chipset and the Radeon HD 6XXX series graphics chipsets. You can compile a custom kernel enabling better radeon support, or just try different kernels available via Fog kernel updates.

working on it now in VirtualBox.

Let me setup another storage node and see what we need. I’m pretty sure you have to sync the usernames and passwords between the storage servers so they can talk, I just don’t know exactly where yet.

There is a great section in the user guide about uploading your first image.

One big different I have is that each model of computer we have has a complete and separate image for it. Like we have a lab of 42 computers, all identical hardware. They have 1 image. We have 120 teacher laptops, all identical hardware, they have 1 image.

I think I have about 11 different images. I have not ventured into the hardware independent image yet because each image usually has it’s own hardware and its own software requirements.

Sysprep has not been needed for me to image because I am not trying to clear out all the hardware info from windows before I make the image.

If you continue to have problems with 12.04, I suggest installing 10.04 and Fog 0.32. The combination seems to work really well.

I used sysprep on my Windows 7 laptops that will be frozen with DeepFreeze, but only so I could “copyprofile” some settings over from the administrator profile to the default profile. I did the windows setup, hit the hotkeys to reboot into audit mode, and load/setup the computer how I want. Then I sysprep + copyprofile to overwrite the default profile with my custom settings for background and some app options. Reboot after sysprep and finish installing.

Basically, I don’t sysprep right before uploading the image. With Windows 7 and Fog 0.32, it’s not necessary; Fog now takes care of the FogPrep steps to avoid boot up problems. You only have to sysprep to achieve a specific goal only attainable through sysprep. Booting a clean windows 7 image does not require sysprep/fogprep.

For my labs, which are currently running Windows XP, I do not sysprep. I load 1 machine up with all the software I need, join it to the domain and get the policies applied and all the updates ( have to join to domain to get windows update to work ). Once it’s complete, I remove it from the domain and shutdown. I register the device if not already registered and I schedule an upload task. Once it’s complete, I can use the image to push to the rest of the lab and Fog will rename each client based on the name it was registered with. If I have Active Directory integration enabled, it will join them back to the domain for me after they are renamed.

Once they show they are joined to the domain, I freeze them using deep freeze console or manually logging into each one if I don’t have access to the console.