To Reproduce Steps to reproduce the behavior:

Deploy a node, go to undeploy a node with configuration drive data.

Expected behavior

• Ultimately, what is expected is that the IPA ramdisk would not attempt to load any configuration when the PXE boot operation was used, but after an upstream change was made to enable "virtual media/dhcpless" cases by default, an outcome of the PTG, one of the Telco platform folks doing other work observed a behavior change which they reported to us upsream.
• Investigation of the logs revealed the "fail safe" nature of the enforcement script didn't execute as expected and required some additional modification. Luckilly, this was only in place for a couple of weeks, i.e. merging right before Thanksgiving and discovered ~2 weeks later.

Bug impact

• The IPA ramdisk could configure itself with the wrong networking configuration utilizing glean.
• This could result in failures for future cleaning operations. It was not detected in CI in part because one of the two jobs which should have detected it was marked non-voting upstream, and was failing. That job should have highlighted some of the issue, while other jobs are not fully exercised in the exact way the reporter indicated.

Known workaround

• Update diskimage-builder to latest upstream version
• Upstream has already reverted out the offending element and is working on correcting the issue.