it's about time that we ditched legacy facts :D

newest voxupupli-test brings a lonter that forces you to use new facts :p

is there a way to trigger a 'reload' on a service instead of 'restart' via a notify?

you could override the restart command with something that reloads I think, but I'm not sure what the implications are

for example, I'm not sure what happens if the service is stopped and receives a refresh; does it then also call restart or start?

the binary/daemon for which the config is changing support a live-reload at runtime via kill -HIP $PID and that is also exposed via ExecReload in the systemd service

there is systemctl (try-)reload-or-restart $unit, which could work

we hacked something like that for consul https://github.com/voxpupuli/puppet-consul/blob/master/manifests/reload_service.pp

if I also do ensure => running on that service, should I just create an Exec resource that does the try-reload-or-restart and is applied via resource ordering after the service resource with ensure => running?

and a different approach for bird https://github.com/voxpupuli/puppet-bird/blob/master/manifests/init.pp#L160

oh

puppetlabs-postgresql also has something similar: reloads when it can, but restarts when it needs to

See the

puppetlabs-postgresql

module at https://forge.puppet.com/puppetlabs/postgresql?src=slack&channel=voxpupuli

I can just override the

restart

property of the service resource?

o/

Can I get reviews for:

hi, i was just wondering if anyone could offer some advice on getting the spec tests to run on https://github.com/voxpupuli/puppet-splunk. I've got some really minor bug fixes to contribute, but I cant get the rspec tests to complete successfully (even without my modifications).

hi, i was just wondering if anyone could offer some advice on getting the spec tests to run on https://github.com/voxpupuli/puppet-splunk. I've got some really minor bug fixes to contribute, but I cant get the rspec tests to complete successfully (even without my modifications).

Failures: 1) splunk::enterprise supported operating systems on solaris-10-i86pc splunk when including forwarder and enterprise is expected to fail to compile and raise an error matching /Do not include splunk::forwarder on the same node as splunk::enterprise/ Failure/Error: it { is_expected.to compile.and_raise_error(%r{Do not include splunk::forwarder on the same node as splunk::enterprise}) } error during compilation: Evaluation Error: Error while evaluating a Resource Statement, Could not autoload puppet/type/archive: Could not autoload puppet/provider/archive/curl: Could not autoload puppet/provider/archive/ruby: Could not find a directory environment named 'production' anywhere in the path: /home/vagrant/puppet-splunk/spec. Does the directory exist? (file: /home/vagrant/puppet-splunk/spec/fixtures/modules/splunk/manifests/forwarder/install.pp, line: 17, column: 5) on node vagrant-centos7.csu.edu.au # ./spec/classes/enterprise_spec.rb55in `block (6 levels) in <top (required)>'

On a centos7 vm, after doing: rvm install ruby-3.1.1 git clone https://github.com/voxpupuli/puppet-splunk cd puppet-splunk/ bundle install bundle exec rake spec

On a centos7 vm, after doing: rvm install ruby-3.1.1 git clone https://github.com/voxpupuli/puppet-splunk cd puppet-splunk/ bundle install bundle exec rake spec

This message was deleted.

https://github.com/rubocop/rubocop/pull/10746 :)

Huh. Why is this breaking

:auto_correct

?

The rake task not being safe came up again yesterday, then I noticed there had been this related PR https://github.com/rubocop/rubocop/pull/10709

We started to use the zabbix module recently. So many little things I want to fix. Not allowed to do it on work time, sadly, and not sure when I'll have personal time to be able to.

Is there a linter for

optional with a default value

?

yes there is