Can Fog work without PXE?

[rockdown]

I understand that this can be a client server setup maybe similar to GSS. Our campus has firewalls on each floor of our building and each building has a firewall. I need to work across multiple buildings and PXE is not something our security or networking is going to allow. They also no longer wish to support GSS and is not paying for the new version to support Windows 10. So my question is will this deploy an image to a machine that has the client installed without the use of PXE? The TFTP probably will not be a problem but PXE is definitely blocked.

[Wayne Workman]

FOG 1.3.0 RC-8 and prior uses PXE and iPXE, only. FOG 2.0 will have the ability to be PXE-less but the details of that are still pie in the sky and 2.0 is a year (probably way more) away from release.

Your problems probably lay with DHCP. Where are the DHCP servers? How many? Simply allow TFTP, FTP, and NFS through the firewalls. PXE is not some magical thing. These are just protocols across the line that any firewall worth a dime can allow or block.

Optionally, you can put storage nodes on each subnet to PXE boot from.

[Tom Elliott]

@Wayne-Workman Don’t forget, 1.3.0 may not have a default mechanism to enable “PXE-Less” booting, but it can be done thanks to the great work by @george1421

[Wayne Workman]

@Tom-Elliott Ah that’s right. While it is an option, it’s not scalable. You can’t click “go” from your desk to image 500 computers on the other side of the planet without someone physically touching each one.

@george1421 What you ought to work on is what has been similarly done already with Macs. Make your FOS stick just a partition on a windows drive, and setup dual-boot with windows as default. Have the FOS partition display a menu as always, with the other OS as a boot option.

[george1421]

I guess the first question for the OP is how do you intend to image. What is your process? Do you want to image machines across your campus from a central location or do you require a tech to sit in front of the target computer during imaging? The answer will impact your options.

[rockdown]

@george1421 Maybe both. at first a central location for my labs but i could expand into single targeted machines.

[george1421]

@rockdown OK maybe I should have not been so what if.

If PXE booting is not an option across your network. I see two options for you.

1. We can usb boot into the FOG (FOS engine) as long as between the FOG server and the target computer has the following protocols enabled: NFS, ftp. and http. These are mandatory protocols and realize that nfs has a bunch of sub protocols behind. The tech must sit near the target computer to press the F12 to select the usb boot option and then from the usb boot menu pick image/deploy. You can not do this from across your campus.

2. Create a portable fog server using a laptop. The tech will take that laptop (mobile deployment server) to the remote subnet and plug it in. This process WILL use pxe but only on that subnet and while the mobile deployment server is connected. The FOG mobile deployment server will have dnsmasq running to supply the missing pxe information required to pxe boot the target computers. With the mobile deployment server connected to the local subnet the IT tech will need to press F12 to pxe boot the target computer into the FOG menu. From there the Tech can register the system and deploy the image.

Either way its not ideal but it is possible. It would be better if you could just get the networking group to allow pxe booting to only your fog server so you can use the tool like it was intended.

[Wayne Workman]

Storage nodes per-building is also an option.

If you want to go with a mobile fog laptop server using dnsmasq, I actually have a project that can automate this for you.