Automatically test the sanity of the systemd journal and syslog

For instance, we could boot Tails/Jessie, grab these logs, go through them line-by-line and create a “canonical” log pattern that we expect to get (no more, no less) at each boot. Obviously we’d strip time stamps and other stuff that can varies, and we would make the order of the entries not matter, etc etc. The test would fail if we see anything unexpected, or miss anything expected. If this is too hard, we could could limit it to “only allow the expected lines with the word ‘error’ in it” and “these lines must be present”.

This could be made into a new step, and/or as a part of the Tails seems to have booted normally step.

Parent Task: #10277

Related issues

• Blocked by #6171 (closed)

Original created by @anonym on 10279 (Redmine)