@Tom-Elliott I understand that the list of packages depends on the OS installed, and, if I understood you correctly, the packages listed in the .fogsettings.

What I’m asking for is a way to calculate, on the target machine, the list of packages (& versions?) that would be needed to be installed in order for fog to be installed for the first time or for the existing fog installation to be upgraded. That’s why I suggested that it be an argument to installfog.sh.

Imagine my fog server is on a network that is completely isolated from the Internet. I have to be able to calculate the list of packages (& versions?) fog needs to operate, cp that list to an SD card or whatever as, say, deps.txt, then walk over to the machine with the same OS that does have Internet, mount the SD card, then do something like cat /media/.../deps.txt | apt-get install --download-only --bundle or whatever. After all the bundles are downloaded, copy them to the SD card & sneakernet them back to the isolated fog server and install them.

Capice?

FYI filed issue https://github.com/FOGProject/fogproject/issues/576

@Tom-Elliott Will do, however, if the list of packages ever changes, I’d somehow need to know that.

What would be reliable, however, is to add the feature as described above: ./installfog.sh --list-packages. What do you think?

Hi all,

I’m curious if there’s a way to facilitate an offline installation of Fog server. In particular, I’d like to be able to ask the installfog.sh script for all of the packages needed, like ./installfog.sh --list-packages so that I can for it in $(./installfog.sh --list-packages); do sudo apt install --download-only $it; done.

Is there a way to do that now?

Thanks,
Matthew

@Sebastian-Roth Thanks. Worked like a charm. Went from ~250Gb image down to ~9Gb.

Hi all,

Note: I’m a fog noob.

Using fog server to PXE boot ubuntu server onto PCs. I noticed that when capturing an image of my base ubuntu server, the image ended up being quite large because of the 256Gb SSD that I installed the OS on. This is basically a vanilla ubuntu server OS with no additional software on it, so it seems to me that most of the space on the /dev/sda3 (the largest partition) is empty. I’m noticing that it’s taking a while to transfer the image over the LAN, and I’d like to speed things up.

Is there a way to have fog reduce the size of the image, given that so much of it is empty space?

If not, what tips or tricks do you suggest to achieve the same result? Like maybe, before capturing the image, resizing the main partition from 250-something Gb down to the smallest size for the OS to fit on?

Also, if I do that, then deploy the image to a PC with, say, a 1 Tb drive, will I have to resize the main partition after deployment, or can I tell fog to use all available space?

Thanks,
Matthew