Fix default value for $service_status on ArchLinux

[smortex]

The value can be overriden from postgresql::globals, but if no
override is provided by the user, the module should use a valid default.
ArchLinux use systemd, so use the relevant systemctl command.

Fixes #1408

[puppet-community-rangefinder]

postgresql::params is a class

Breaking changes to this file WILL impact these 2 modules (exact match):

• tailoredautomation-patroni
• aursu-gitlabinstall

Breaking changes to this file MAY impact these 3 modules (near match):

• maestrodev-continuum
• puppetfinland-bacula
• puppetfinland-trac

This module is declared in 70 of 580 indexed public Puppetfiles.

---

These results were generated with Rangefinder, a tool that helps predict the downstream impact of breaking changes to elements used in Puppet modules. You can run this on the command line to get a full report.

Exact matches are those that we can positively identify via namespace and the declaring modules' metadata. Non-namespaced items, such as Puppet 3.x functions, will always be reported as near matches only.

[ekohl]

Isn't it better to change the type definition to allow Undef? Then the service provider can figure it out by itself.

Actually something we should do for most OSes anyway instead of writing out the systemd command.

[smortex]

That make sense: this module has a load of legacy and I trend to contribute to it by copying the patterns I see in the other OS code and are missing rather than killing everything with fire and make everything simple. I'll work on it on another PR.

[smortex]

Ah, these commands are used by the "instance" reloading code (when a single service manage multiple clusters). This rely on an exec resource that run these commands, so they can't be undef and must be explicitly set ATM.

I guess the initial idea was to allow to restart a single cluster (i.e. pg_ctlcluster 11 main restart on debian) but may have not been fully implemented (yet)?

[ekohl]

@SimonHoenscheid is working on that, but this isn't going to work when it's defined in the class. Perhaps it needs to be modified to something interpolated, though a safe implementation would use an array. I wonder what's best to do here.

[smortex]

While the PR is not perfect, it can unbreak the module for ArchLinux and that would unbreak CI for some of our Voxpupuli modules. Maybe this fix can go as it is for now (and be released as a patch version), and updated again when we have some cluster management code which would be at least a minor version bump?

[genebean]

This seems inline with other code within the file. Refactoring the process seems like something for a different PR to me.

[ekohl]

Fair enough.

[smortex]

Rebased on top of main so that the branch does not appear out-of-date for GitHub (Is that a repo requirement or maybe I got fooled by GitHub, this nuked all reviews 😒)?

[Phil-Friderici]

@puppetlabs/modules any chance to review this very tiny PR that is only one line and would help me/us greatly ?

[bastelfreak]

I rebased it against main to fix the SLES errors.

[bastelfreak]

as mentioned in the comments, that's not perfect but will fix Arch Linux.

[Phil-Friderici]

@bastelfreak THANK YOU, Thank You, thank you