Created attachment 920650 [details] debug log Description of problem: setup a external capsule and included the subnet 192.168.200.1/24 in the host build. when creating the host the following prevents from saving: Failed to generate PXELinux template: undefined method `each' for nil:NilClass This seemed to pop up when adding subnet to the host group. I am including production.log with debug turned on
Since this issue was entered in Red Hat Bugzilla, the release flag has been set to ? to ensure that it is properly evaluated for this release.
additional info...the nilClass issue only happens when I attach a subnet to the Host Group or select manually for each host build.
Hey Shannon, we do not propagate the exception properly, can you please edit this file: /usr/share/foreman/app/models/concerns/orchestration/tftp.rb and on the line 82 81 rescue => e 82 failure _("Failed to...") % { :template_kind => os.template_kind, :e => e } 83 end add extra parameter e as the second one: 82 failure _("Failed to...") % { :template_kind => os.template_kind, :e => e }, e Then reproduce and paste the exception backtrace. Note for assignee: we should fix all "failure" calls when exception is not passed in as a second paramter with this bug.
attaching stack as stack.trace
Created attachment 921761 [details] stack trace for tftp
Sorry, the line should be: 82 failure _("Failed to...") % { :template_kind => os.template_kind, :e => e }, e.backtrace (note the "backtrace" method call). This stracktrace is useless :-(
It's likely the wrong PXELinux template is associated to the operating system. Ensure you have "Kickstart default PXELinux" associated, *not* "PXELinux global default" or similar.
Created upstream ticket - let's improve error handling for 6.1+. http://projects.theforeman.org/issues/7258
Can't reproduce the original issue in this ticket. Error handling/logging has a dedicated BZ: https://bugzilla.redhat.com/show_bug.cgi?id=1194352
Moving to POST since upstream bug http://projects.theforeman.org/issues/7258 has been closed
I have been able to provision a new host (libvirt installed in the Satellite itself) using an external capsule (complete different system) without any issues. I used the network/subnet associated with my libvirt setup and everything worked fine.
Spoke to @shugues and @Toledo about this issue and we all feel that it is safe to close it as WORKSFORME.