Plugin Hooks Not Running at Sub-Site
-
Hello everyone,
I have an interesting issue I am facing. I am testing having a storage node at another physical site. The connection seems fine as it can reach the FOG server but one thing I am facing is I am heavily reliant on the Subnet Groups plugin, specifically for identifying what IP a client is upon booting into FOG and assigning to the associated group.
What I am finding is that when a client boots at the sub-site it will get the boot files from the storage node and get into the ‘DEVICE IS REGISTERED’ screen but the plugin hook never fires thus thwarting my desired process.
My desired process is when computers are setup in a lab: Booted > Check-in with FOG > Subnet Group assigns computers to proper group > boot loop until all devices are in group > assign an image and task > machines are images. I have this working at the main site.
Is my best option to have a separate full FOG setup at each site?
Appreciate anyone’s input!
FOG version: 1.5.10.1733
Server OS: Ubuntu 24.04 -
To follow up on this I am looking for any community input on the following:
If anyone has a multi-site setup where they are reliant on the Subnet Group plugin did you:
- Edit the init.xz file to check in with the master server, if so, I’m curious how
- Setup each site node as a normal server with your full configuration
- Another option that I have not thought of
Anyone’s input would be greatly appreciated.
-
@christop When I was using fog I worked at a company that had many sites. There was a master fog server at HQ and storage nodes at each site. I don’t know what the subnet group plugin is so I can’t give direction there. But for us we had one image for global deployment. Then we used a fog post install script that updated the target computer’s unattend.xml file at imaging time. This allowed us to update the system’s default keyboard, locale, language, and destination OU the system was assigned to. We did not use FOG snapins for deployment but another tool once imaging is done. This other solution was used to deploy applications that could not be baked into the golden image because they had deploy time IDs that we did not want replicated to all deployed systems (like Antivirus system IDs).
If you are interested in this method I do have some tutorials that will give you a head start. But if you are using this subnet group assignments to add the machines to a snapin deployment group then my method will not work well.
-
@george1421 thanks for the detailed response! We are frequently re-imaging lab rooms. The subnet group plugin is vital as when a device moves rooms to a separate subnet, when it ‘checks-in’ to FOG after it is powered on and PXE boots in the new location it will be re-assigned to the proper group associated with the IP range. That subnet group will be used to group all the lab computers in the desired room and thus used to multicast the room.
The issue lies when I have a storage node setup at a new location and acting as the PXE/TFTP server, the client will not ‘check-in’ with the main FOG server and run the plugin hooks until it finishes imaging and the client takes care of it which does not work in my desired process.
At this point I am leaning towards a standalone server for each site just to support the functionality required.
-
@christop said in Plugin Hooks Not Running at Sub-Site:
lab computers in the desired room and thus used to multicast the room
just a data point here. Only the master fog server “the one with the database” can multicast images. Storage nodes can not (unless something changed in the last few years). Storage nodes are basically NAS devices with a little programming.
Now we used the location plugin to create storage locations. We assigned the storage nodes to a location and then target computers to locations, so as the target computer boots it finds the storage node it should image from. But that won’t help with the multicast part because storage nodes can only unicast images.
