Architecture Question

Hi everyone,

I have a ~500 server build that needs to be done, and I’d like to use Foreman to do so. My initial testing went really well and I like how simple Foreman was to setup and provision servers. Now that my testing is done, I had a few questions regarding some of the architecture.

First of all, I have two subnets that I want to use. One subnet is for the servers primary network interface which will also handle provisioning. This is a bonded interface (LACP on the switch), with multiple VLANs tagged. This subnet will have DHCP/TFTP/PXE enabled through the default proxy on the Foreman server.

A second subnet and VLAN will be for the BMC (Dell iDRAC) which I wanted to deploy a separate smart-proxy for to run DHCP/BMC.

The idea is to let Foreman handle all DHCP and discovery of the servers when they come up on the network so it can populate Foreman with all of their facts and information - including their BMC network interfaces that would DHCP separately.

  1. Does this make sense?

  2. Are there any issues using Foreman to provide DHCP for the BMC + the server’s NICs + autodiscovery?

  3. What is the difference between the “Subnets” list in the UI vs. configuring DHCP ranges for the proxies on the command line?

1 Like

Hey and welcome.

Well it does, however if you can stay away from VLAN/bonding for provisioning interfaces you can make your life easier. While this should work, there are often bugs in templates/OS installer networking (dracut/anconda/debian-installer) and you might need to solve some things.

BMC interface for DHCP is like any other device. But for discovery/autodiscovery keep in mind that the features are very limited - there is no mechanism to create bonds from discovered hosts via autodiscovery rules at the moment.

IP management is very simplified in Foreman, range defines what Foreman tries to look for free IP addresses:

https://docs.theforeman.org/guides/build/doc-Provisioning_Guide/index-foreman.html#networking_resources

1 Like