Description of problem: Even after taking care of everything as mentioned in BZ# https://bugzilla.redhat.com/show_bug.cgi?id=2170908 , from Capsule's perspective, The Discovery fails as The host tries to get discovered via Satellite always. Version-Release number of selected component (if applicable): # rpm -q satellite rubygem-foreman_discovery satellite-6.13.0-6.el8sat.noarch rubygem-foreman_discovery-22.0.2-1.1.el8sat.noarch How reproducible: Always Steps to Reproduce: 1. Setup a Satellite 6.13 and Capsule 6.13 on libvirt 2. Configure the Capsule server with DHCP\DNS\TFTP\Discovery_via_iPXE and configure the satellite to support the capsule for deployments 3. Ensure that suggestions from https://bugzilla.redhat.com/show_bug.cgi?id=2170908 have been taken into consideration. 4. Try booting a blank system into the same network as the capsule so Discovery can be attempted via the capsule's network Actual results: The Discovery will be attempted but once the fdi vmlinuz and initrd files are loaded via httpboot endpoints from capsule, It seems the process further renders the "iPXE Global Default" template and then gets the satellite URL as the discovery endpoint ( as it's hardcoded in the template ). Expected results: There should be some way present for the Discovery to work from any external capsules. Additional info: NA
Is this a regression from 6.12? Thanks!
Considering the workflow it follows, I don't think this is a regression but It has never really worked. Due to a lack of testing with iPXE, This issue was never captured properly but very recently ( in the past year ), we have seen users trying to make use of iPXE in various ways ( from satellite as well as capsules ) and hence the effort to identify what it can do and what it can't. If we can ensure that our documentation is clear enough to explain how far iPXE can work and under which specific circumstances it won't, That would help the end users greatly.