Bugfixes for firstboot script issues

[JedMeister]

Built on top of #66 & #67 - these PRs can be independently reviewed separately/first or just review everything in this PR if preferred - whatever works best for you.

Fixes:

• c3531aa - firstboot.d/15regen-sslcert was failing due to nginx.service being listed by systemd but didn't actually exist
  • now only attempts to restart services which are actually running; rationale is that services which aren't running don't need to be restarted as restart is only to support updated ssl certs.
• 2bdd929 - firstboot.d/29tagid was failing due to initfence html file not existing in /var/lib (due to package refactoring in Update turnkey init fence #66)
  • now if /var/lib/inithooks/turnkey-init-fence dir does not exist, it is copied from /usr (where it was moved to in e77d288) before the changes are applied; rationale is that the files that are included with the package should be in /usr/lib (as per packaging change noted above) but files in /usr/lib should not be modified, instead the modified files should be in /var/lib - thus need to be copied across if they don't exist.

Further testing is still required before we can consider this "done".

During preliminary testing it appears that the original changes in #67 (updated mechanism to launch inithooks/confconsole on tty1) still don't work quite as intended - at least not on the v19.0beta core I'm running live. I'm not sure why, but I manually started the new inithooks.service and it hasn't rerun since then. It could be a race condition during boot perhaps? Or perhaps there is something that we haven't cleaned up properly/completely from the old way we used to do things?

---

[update] Pushed new commit: 1d894b6. I'm not sure if that is the actual fix for the (ongoing) gett1 issue noted directly above, however it was clearly a bug.

Also I've noticed that my local trixie testing ISO build still has these files:

• /etc/systemd/system/getty@tty1.service.d/10-getty-tkl-login.conf
• /etc/systemd/system/getty@tty1.service.d/override.conf
• /usr/lib/inithooks/bin/login_script.sh

Besides the fact that they shouldn't be there anymore at all (after #66 & #67) it's a bit weird that they ever existed together because both of the getty overrides do the same thing?! I.e. both override getty1 to start inithooks (prior bookworm functionality). Obviously I'm missing something but after checking the package contents (i.e. the code here) and common, I can't see where they are coming from?!

[JedMeister]

bump @OnGle

[OnGle]

lgtm