Quadlet - Allow setting Service WorkingDirectory for Kube units

[ygalblum]

Add key for Quadlet to set WorkingDirectory to the directory of the YAML or Unit file Add Doc
Add E2E tests
Add System test

Does this PR introduce a user-facing change?

Yes

Quadlet - Allow the user to instruct Quadlet to set the service working directory relative to the YAML or Unit files

Resolves #17177

[Luap99]

I don't understand the use case, why do we need a new field for that?

A user can already set:

[Service]
WorkingDirectory=/mypath

What am I missing?

[ygalblum]

What am I missing?

The issue is how transferable the unit file is. For example, let's say that we have a service that different users want to run. They can each place it under /etc/containers/systemd/users/$UID. Now, assuming they want the path to point into this directory, only the generator can handle it automatically. Your suggestion will require setting this field differently for each user.

[Luap99]

Right but for that you can use the systemd specifiers (%t,%h,etc...)
I just wonder if this is really needed? But if you think that makes in easier to use for users I have no objections.

code LGTM

[ygalblum]

Maybe the specifiers can solve the use case I added. But, from the discussion in #17177 it seems that having Quadlet resolve the paths is desired

[rhatdan]

I think the issue right now is that this is random, and not something the user can easily expect.

[rhatdan]

What is the default for this now? I am more interested in the default WorkingDir be set to the location of the quadlet.

[Luap99]

Systemd defaults to the users $HOME when rootless and to / when run as root AFAICT.

[vrothberg]

LGTM, nice work!

[edsantiago]

Not a blocker (right now), but please join the discussion here: #19236 (review)

[TomSweeneyRedHat]

@ygalblum some of the tests are still a bit red.

[edsantiago]

Test failures were podman search, one of our most popular flakes. Passed on rerun.

@ygalblum if you could kindly update your system test to the non-journalctl form I'm using in #19276 I would be most grateful, but I will also understand if you've had quite enough of the back-and-forth on this.

[ygalblum]

you could kindly update your system test

@edsantiago I've updated my test to not use journalctl the same way as in #19284

[edsantiago]

Thank you! LGTM but I will let others make the final approval.

[vrothberg]

/lgtm

[openshift-ci]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: edsantiago, vrothberg, ygalblum

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [edsantiago,vrothberg,ygalblum]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment