Remove support for Debian 9 and Ubuntu 16.04; Add Debian 10 support

[wyardley]

Pull Request (PR) description

• Deprecate Debian 9 and Ubuntu 16.04
• Add support for Debian 10
• Minor test updates (remove RABBITMQ_USE_LONGNAME in clustering acceptance tests)
• Fix rabbitmq_plugin to correctly detect implicitly enabled plugins (Cherry-picked from Fix rabbitmq_plugin to correctly detect implicitly enabled plugins #844)

This Pull Request (PR) fixes the following issues

Closes #844
Fixes #930
Fixes #887

[wyardley]

@bastelfreak actually, this is not as bad as I thought. acceptance tests basically pass - I can see if I can figure out an explanation / fix for the one failing test.
Let me know if there's a better way to handle the title / description for both removing and adding support for versions in the same PR, or if it needs to be done in two phases.

  1) rabbitmq parameter on a vhost: create parameter resource runs successfully
     Failure/Error: apply_manifest(pp, catch_changes: true)
     Beaker::Host::CommandFailure:
       Host 'debian10-64-puppet7.example.com' exited with 2 running:
        puppet apply --verbose --detailed-exitcodes /tmp/apply_manifest_195729764.pp.4YjFap
       Last 10 lines of output were:
       	Info: Loading facts
       	Info: Loading facts
       	Notice: Compiled catalog for debian10-64-puppet7.example.com in environment production in 0.29 seconds
       	Info: Using environment 'production'
       	Info: Applying configuration version '1683748656'
       	Notice: /Stage[main]/Main/Rabbitmq_plugin[rabbitmq_federation]/ensure: created
       	Info: /Stage[main]/Main/Rabbitmq_plugin[rabbitmq_federation]: Scheduling refresh of Service[rabbitmq-server]
       	Notice: /Stage[main]/Rabbitmq::Service/Service[rabbitmq-server]: Triggered 'refresh' from 1 event
       	Info: Calling prefetch: documentumFed@fedhost
       	Notice: Applied catalog in 18.74 seconds

[bastelfreak]

ah nice, thanks! I think the title is fine.

[wyardley]

  Notice: Applied catalog in 61.59 seconds

debian10-64-1 executed in 89.26 seconds
Exited: 2
    runs successfully (FAILED - 3)

debian10-64-1 13:50:34$ rabbitmqctl list_parameters -p fedhost
  Listing runtime parameters for vhost "fedhost" ...
  federation-upstream	documentumFed	{"expires":3600000,"uri":"amqp://server"}

debian10-64-1 executed in 2.40 seconds
    has the parameter

So, if I run the manifest again locally, the exit code is 0.
Also, the actual test that the parameter exists is there.

Not sure if it's just an idempotency or timing issue? There's no output that suggests an actual problem.

[wyardley]

Worse, it doesn't seem to be totally consistent what fails 🤦

[wyardley]

Ubuntu 22 has a lot more failures; going to leave that out for now.
If anyone on IRC (or any folks who have been more involved recently with Vox / Puppet stuff) has time to try and fix modulesync, that would also be super helpful, because I am wondering if that might help make sure all the integration test scaffolding is up to date.

[wyardley]

Ok, pushed a test fix for some clustering related stuff.

The other stubborn one is the "create parameter resource" one, where the test itself works fine, but the idempotency check (no changes on second run) fails. Going to look at that one.

  Info: Applying configuration version '1683775408'
  Notice: /Stage[main]/Main/Rabbitmq_plugin[rabbitmq_federation]/ensure: created

The issue w/ Ubuntu 20.04 seems to be related to Python installation seemingly needing to run each time.

[wyardley]

It looks like the main issue is due to the newer RabbitMQ version and the issue outlined in #844

[wyardley]

I gave up on Debian 11 / Ubuntu 20 for today. Someone can sort those out after some of the rest of this gets out.

Arch failures were preexisting.