https://www.puppet.com/community logo
Join Slack
Powered by
# voxpupuli
  • p

    Pat Riehecky

    03/17/2023, 2:15 PM
    My usual selection mode is: is there a "supported" module, is there a module provided by the thing I'm managing, is there a voxpupli module, is there an "approved" module, is there a pdk module
  • c

    CVQuesty

    03/17/2023, 2:16 PM
    i’m the same. just move PDK up to 2nd in that list
  • v

    VoxBot

    03/17/2023, 2:19 PM
    well, https://github.com/voxpupuli/beaker-rspec/pull/120 now passes on Ruby < 3.2 so that does mean better test coverage :)
  • c

    CVQuesty

    03/17/2023, 2:19 PM
    Now that I think about it, it’s supported, approved, pdk, Vox. Outside of those things, I try not to use random modules on the Forge. I do know that Perforce is trying to clean up the Forge now. Waiting to see what that looks like.
  • c

    CVQuesty

    03/17/2023, 2:20 PM
    And if the testing framework finally gets settled, and they make it a hard requirement, I’m sure there’ll be quite a bit of “scorched earth” at some point.
  • v

    VoxBot

    03/17/2023, 2:21 PM
    we've briefly chatted about Litmus vs Beaker at cfgmgmtcamp and I believe it should be fairly easy to align the two so that you can use the same syntax for both and they're really just libraries
  • v

    VoxBot

    03/17/2023, 2:22 PM
    my idea is that you start using native rspec matchers instead of apply_manifest() - so https://github.com/voxpupuli/beaker-rspec/pull/115 is roughly where I'd like to end up
  • v

    VoxBot

    03/17/2023, 2:22 PM
    and then write the same matcher but with a Litmus implementation
  • v

    VoxBot

    03/17/2023, 2:24 PM
    with that I think it'd be much smoother for users if PDK does decide to move back to Beaker
  • c

    CVQuesty

    03/17/2023, 2:26 PM
    that’s a great idea.
  • t

    tvaughan

    03/17/2023, 2:29 PM
    As long as you stay away from
    serverspec
    , you'll be OK with that.
  • c

    CVQuesty

    03/17/2023, 2:30 PM
    I was wondering when you were going to wake up and weigh in here. 😛
  • v

    VoxBot

    03/17/2023, 2:31 PM
    we do heavily rely on serverspec today, but I've wondered about multi server: I have seen an
    on: host
    parameter somewhere and wonder if that's actually implemented
  • v

    VoxBot

    03/17/2023, 2:32 PM
    I've thought about alternatives, because I do like some of the abstractions that it provides
  • t

    treydock

    03/17/2023, 2:45 PM
    I've used the multi server setup a lot with beaker , never could get it to work with Litmus but those attempts were a year or two ago
  • t

    treydock

    03/17/2023, 2:48 PM
    https://github.com/sensu/sensu-puppet/blob/master/spec/acceptance/00_backend_spec.rb - the majority of Sensu acceptance tests run a backend and agent as two distinct containers and run tests differently between the two including running specific serverspec tests on different containers.
  • t

    treydock

    03/17/2023, 2:48 PM
    https://github.com/sensu/sensu-puppet/blob/master/spec/acceptance/00_backend_spec.rb - the majority of Sensu acceptance tests run a backend and agent as two distinct containers and run tests differently between the two including running specific serverspec tests on different containers.
  • v

    VoxBot

    03/17/2023, 2:58 PM
    ewoud: we do some multinode testing here: https://github.com/voxpupuli/puppet-openvpn/blob/master/spec/acceptance/openvpn_spec.rb
  • v

    VoxBot

    03/17/2023, 3:02 PM
    bastelfreak: yes, but I get the impression that all the assertions are on the server node
  • v

    VoxBot

    03/17/2023, 3:02 PM
    still, good place to start to improve things
  • v

    VoxBot

    03/17/2023, 3:04 PM
    yeah true
  • v

    VoxBot

    03/17/2023, 3:04 PM
    https://github.com/voxpupuli/beaker/pull/1784 mhm I think it fails to provision the container
  • v

    VoxBot

    03/17/2023, 3:12 PM
    ewoud: oh maybe you need to switch to the ubuntu20.04 runner, like we did it in gha-puppet
  • v

    VoxBot

    03/17/2023, 3:12 PM
    or don't use centos7
  • v

    VoxBot

    03/17/2023, 3:15 PM
    before the sshd patch it worked on Ruby < 3.2
  • d

    David Sandilands

    03/17/2023, 3:15 PM
    I think it reflects a past period where Puppet org was not communicating and developing together with Vox properly so to be fair i think divergence was inevitable. While there will be always be some difference in approach hopefully what we should be seeing out of the new ecosystem product team is a far greater focus to narrow these differences and reduce duplication of work by working together.
  • v

    VoxBot

    03/17/2023, 3:15 PM
    bastelfreak: so perhaps the sshd bit actually breaks centos 7?
  • v

    VoxBot

    03/17/2023, 3:15 PM
    mhm yeah possible
  • v

    VoxBot

    03/17/2023, 3:23 PM
    ewoud: did you test it before the sshd patch?
  • v

    VoxBot

    03/17/2023, 3:23 PM
    bastelfreak: yes, see the commit before last
1...367368369...648Latest