tests: add test of systemctl restarting snapd in various ways

[olivercalder]

There was a potential concern that fdstore changes from #16119 could affect snapd restarts.

In particular, I had observed that the following was blocking while I was working on #15508:

systemctl restart snapd.service snapd.socket

and anecdotally, the following worked instead:

systemctl restart snapd

However, I now believe that was caused by a kernel bug with notification protocol v5, rather than a problem with the fdstore PR #16119

Therefore, I tentatively believe the PR to revert the fdstore changes is not necessary: #16228

This work is tracked (in part) by https://warthogs.atlassian.net/browse/SNAPDENG-36554

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 78.92%. Comparing base (89c2bd8) to head (cd6fc24).

Additional details and impacted files

@@            Coverage Diff             @@
##           master   #16231      +/-   ##
==========================================
+ Coverage   78.86%   78.92%   +0.05%     
==========================================
  Files        1403     1396       -7     
  Lines      196311   196092     -219     
  Branches     2462     2462              
==========================================
- Hits       154829   154766      -63     
+ Misses      32190    32039     -151     
+ Partials     9292     9287       -5     

Flag	Coverage Δ
unittests	78.92% <ø> (+0.05%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[github-actions]

Wed Jul 22 18:40:51 UTC 2026
The following results are from: https://github.com/canonical/snapd/actions/runs/29944385520

Failures:

Preparing:

• openstack:opensuse-tumbleweed-64:
• openstack:opensuse-tumbleweed-64:
• openstack:opensuse-tumbleweed-64:
• openstack:opensuse-tumbleweed-64:
• openstack:opensuse-tumbleweed-64:

Executing:

• garden:ubuntu-core-18-64:tests/core/gadget-kernel-refs-update-pc
• openstack:ubuntu-core-18-64:tests/main/systemctl-restart-snapd-variations:restart_snapd_service_socket
• openstack:ubuntu-core-20-64:tests/main/systemctl-restart-snapd-variations:restart_snapd_service_socket
• openstack:ubuntu-core-22-64:tests/main/systemctl-restart-snapd-variations:restart_snapd_service_socket
• openstack:ubuntu-20.04-64:tests/main/systemctl-restart-snapd-variations:restart_snapd_service_socket
• openstack:ubuntu-22.04-64:tests/main/systemctl-restart-snapd-variations:restart_snapd_service_socket_reload

Restoring:

• openstack:opensuse-tumbleweed-64:
• openstack:opensuse-tumbleweed-64:
• openstack:opensuse-tumbleweed-64:
• openstack:opensuse-tumbleweed-64:
• openstack:opensuse-tumbleweed-64:

Skipped tests from snapd-testing-skip

If you wish to have any of the below tests run in your PR, in your PR description, add 'unskip:' followed by a copy-and-pasted list of the below tests you wish to run (unskip plus test list must be valid yaml)

• garden:ubuntu-25.10-64:tests/main/apparmor-prompting-support

[olivercalder]

Nevermind, some tests failed! For example: https://github.com/canonical/snapd/actions/runs/18980154198/job/54211757242

Let's see if the same thing occurs with the fdstore changes reverted...

[ZeyadYasser]

@olivercalder should this PR be closed or rebased?

[olivercalder]

Rebased to see if tests are happy

[olivercalder]

I rebased, ran this locally, and everything passed on ubuntu-{20,22,24,26}.04-64. So I don't think there's anything interesting to learn here by having this in the test suite. The aforementioned prompting kernel bug only occurs if prompting is actually enabled, which it's not here. So probably not worth adding this test to run everywhere when there should not be issues with restarting snapd.