Restart PostgreSQL when quadlet env vars change

[shubhamsg199]

Why are you introducing these changes? (Problem description, related links)

The Deploy PostgreSQL container task updates the quadlet file with tuning-specific env vars (POSTGRESQL_MAX_CONNECTIONS, POSTGRESQL_SHARED_BUFFERS, POSTGRESQL_EFFECTIVE_CACHE_SIZE) but does not notify the Restart postgresql handler. As a result, switching tuning profiles (e.g. medium to default) updates the quadlet file and triggers daemon_reload, but PostgreSQL keeps running with stale env vars from the previous profile.

What are the changes introduced in this pull request?

• Adding notify: Restart postgresql to the container task ensures the handler fires whenever the quadlet content changes, restarting the service with the updated env vars.

How to test this pull request

Steps to reproduce:

• Run foremanctl deploy --tuning medium
• Check the postgresql tuning values
• Run foremanctl deploy --reset-tuning
• Check the postgresql tuning values again
• The tuning values will be for the medium profile

Checklist

• Tests added/updated (if applicable)
• Documentation updated (if applicable)

[ehelms]

This reads like we would benefit from a test that switches tuning profiles and checks that the values for said tuning profile are updated within the running container(s) that it applies to.

[evgeni]

This reads like we would benefit from a test that switches tuning profiles and checks that the values for said tuning profile are updated within the running container(s) that it applies to.

That's what SatelliteQE/robottelo#21357 is about where this was found, yeah.

I am not sure we want to test every permutation of parameters in our primary CI, as that will be expensive (time wise)

[evgeni]

Tested with foremanctl-2.0.0-1.20260428184728543158.pr479.12.ge60ca68.el9.noarch, PostgreSQL is correctly restarted and the new max_connections value is used.