Thanks for the report, can you confirm how long it took from the errors to the bootstrap VM disk running out of space? The disk space is finite, and the VM is expected to be ephemeral, so I think it's reasonable to expect users to retrieve logs quite soon after any failure - is this a case of the disk space running out soon after failure, or after many hours in an errored state?
This bug was opened against a very old version, I'm closing it. If you still experience the issue, please open a bug (in jira) against a new version with updated information. P.S. We've recently fixed an issue with Ironic keeping its logs locally in addition to the normal podman logs. That could be it.