provision.yaml expansion

[cdenneen]

Currently you can specify a provision.yaml in order to provision multiple images.
What would be useful is to add a collection key to the inventory of hosts so if you just run rake "litmus:install_agent" it would reference this provision file for a possible collection in order to provide different agent's across the inventory?

Current:

---
centos7-puppet6:
  provisioner: docker
  images:
  - litmusimage/centos:7
centos7-puppet7:
  provisioner: docker
  images:
  - litmusimage/centos:7

Proposal:

centos7-puppet6:
  provisioner: docker
  images:
  - litmusimage/centos:7
  collection: puppet6
centos7-puppet7:
  provisioner: docker
  images:
  - litmusimage/centos:7
  collection: puppet7

Currently in order to accomplish this you must run:
litmus:provision_list[centos7-puppet6] followed by litmus:agent_install[puppet6,localhost:2222]
and then
litmus:provision_list[centos7-puppet6] followed bylitmus:agent_install[puppet7,localhost:2223]
obviously with the target port changing this is difficult from an automation perspective

The proposal would be that by default if no key is specified for litmus:provision_list to loop through them all as well as the same loop when it comes to litmus:install_agent ... currently both of these without keys error:

provision_list:

❯ PDK_PUPPET_VERSION=6 pdk bundle exec rake litmus:provision_list
pdk (INFO): Using Ruby 2.5.8
pdk (INFO): Using Puppet 6.19.1
rake aborted!
No key  in ./provision.yaml, see https://puppetlabs.github.io/litmus/Litmus-core-commands.html#provisioning-via-yaml for examples
/Users/cdenneen/.pdk/cache/ruby/2.5.0/gems/puppet_litmus-0.24.0/lib/puppet_litmus/rake_tasks.rb:28:in `block (2 levels) in <top (required)>'
/opt/puppetlabs/pdk/share/cache/ruby/2.5.0/gems/honeycomb-beeline-2.4.0/lib/honeycomb/integrations/rake.rb:21:in `block in execute'
/opt/puppetlabs/pdk/share/cache/ruby/2.5.0/gems/honeycomb-beeline-2.4.0/lib/honeycomb/client.rb:70:in `start_span'
/opt/puppetlabs/pdk/share/cache/ruby/2.5.0/gems/honeycomb-beeline-2.4.0/lib/honeycomb/integrations/rake.rb:16:in `execute'
/opt/puppetlabs/pdk/share/cache/ruby/2.5.0/gems/bundler-2.1.4/lib/bundler/cli/exec.rb:63:in `load'
/opt/puppetlabs/pdk/share/cache/ruby/2.5.0/gems/bundler-2.1.4/lib/bundler/cli/exec.rb:63:in `kernel_load'
/opt/puppetlabs/pdk/share/cache/ruby/2.5.0/gems/bundler-2.1.4/lib/bundler/cli/exec.rb:28:in `run'
/opt/puppetlabs/pdk/share/cache/ruby/2.5.0/gems/bundler-2.1.4/lib/bundler/cli.rb:476:in `exec'
/opt/puppetlabs/pdk/share/cache/ruby/2.5.0/gems/bundler-2.1.4/lib/bundler/vendor/thor/lib/thor/command.rb:27:in `run'
/opt/puppetlabs/pdk/share/cache/ruby/2.5.0/gems/bundler-2.1.4/lib/bundler/vendor/thor/lib/thor/invocation.rb:127:in `invoke_command'
/opt/puppetlabs/pdk/share/cache/ruby/2.5.0/gems/bundler-2.1.4/lib/bundler/vendor/thor/lib/thor.rb:399:in `dispatch'
/opt/puppetlabs/pdk/share/cache/ruby/2.5.0/gems/bundler-2.1.4/lib/bundler/cli.rb:30:in `dispatch'
/opt/puppetlabs/pdk/share/cache/ruby/2.5.0/gems/bundler-2.1.4/lib/bundler/vendor/thor/lib/thor/base.rb:476:in `start'
/opt/puppetlabs/pdk/share/cache/ruby/2.5.0/gems/bundler-2.1.4/lib/bundler/cli.rb:24:in `start'
/opt/puppetlabs/pdk/share/cache/ruby/2.5.0/gems/bundler-2.1.4/exe/bundle:46:in `block in <top (required)>'
/opt/puppetlabs/pdk/share/cache/ruby/2.5.0/gems/bundler-2.1.4/lib/bundler/friendly_errors.rb:123:in `with_friendly_errors'
/opt/puppetlabs/pdk/share/cache/ruby/2.5.0/gems/bundler-2.1.4/exe/bundle:34:in `<top (required)>'
/opt/puppetlabs/pdk/private/ruby/2.5.8/bin/bundle:23:in `load'
/opt/puppetlabs/pdk/private/ruby/2.5.8/bin/bundle:23:in `<main>'
Tasks: TOP => litmus:provision_list

install_agent: (From the error looks like it should be attempting to install the latest agent but fails as it can't find it)

❯ PDK_PUPPET_VERSION=6 pdk bundle exec rake litmus:install_agent
pdk (INFO): Using Ruby 2.5.8
pdk (INFO): Using Puppet 6.19.1
install_agent
rake aborted!
Installation of agent failed.
Results:
  localhost:2225: {"_output"=>"12:26:47 -0500 INFO: Version parameter not defined and no agent detected. Assuming latest.\n12:26:47 -0500 INFO: Downloading Puppet latest for el...\n12:26:47 -0500 INFO: Red hat like platform! Lets get you an RPM...\n12:26:47 -0500 INFO: Downloading http://yum.puppet.com/puppet-release-el-7.noarch.rpm\n12:26:47 -0500 INFO:   to file /tmp/install.sh.86827.58897/puppet-release-el-7.noarch.rpm\n12:26:47 -0500 INFO: Trying curl...\n12:26:47 -0500 INFO: installing puppetlabs yum repo with rpm...\nLoaded plugins: fastestmirror, ovl\nLoading mirror speeds from cached hostfile\n * base: mirror.vtti.vt.edu\n * epel: epel.mirror.constant.com\n * extras: mirror.nodespace.net\n * updates: mirror.ash.fastserv.com\nNo package puppet-agent available.\n12:26:49 -0500 INFO: Retrying... [1/5]\nLoaded plugins: fastestmirror, ovl\nLoading mirror speeds from cached hostfile\n * base: artifactory.ap.org\n * epel: epel.mirror.constant.com\n * extras: artifactory.ap.org\n * updates: artifactory.ap.org\nNo package puppet-agent available.\n12:26:50 -0500 INFO: Return code: 1\n12:26:51 -0500 INFO: Retrying... [2/5]\nLoaded plugins: fastestmirror, ovl\nLoading mirror speeds from cached hostfile\n * base: artifactory.ap.org\n * epel: epel.mirror.constant.com\n * extras: artifactory.ap.org\n * updates: artifactory.ap.org\nNo package puppet-agent available.\n12:26:52 -0500 INFO: Return code: 1\n12:26:53 -0500 INFO: Retrying... [3/5]\nLoaded plugins: fastestmirror, ovl\nLoading mirror speeds from cached hostfile\n * base: artifactory.ap.org\n * epel: mirror.steadfastnet.com\n * extras: artifactory.ap.org\n * updates: artifactory.ap.org\nNo package puppet-agent available.\n12:26:54 -0500 INFO: Return code: 1\n12:26:55 -0500 INFO: Retrying... [4/5]\nLoaded plugins: fastestmirror, ovl\nLoading mirror speeds from cached hostfile\n * base: mirror.vtti.vt.edu\n * epel: d2lzkl7pfhq30w.cloudfront.net\n * extras: artifactory.ap.org\n * updates: artifactory.ap.org\nNo package puppet-agent available.\n12:26:56 -0500 INFO: Return code: 1\n12:26:57 -0500 INFO: Retrying... [5/5]\nLoaded plugins: fastestmirror, ovl\nLoading mirror speeds from cached hostfile\n * base: artifactory.ap.org\n * epel: mirror.csclub.uwaterloo.ca\n * extras: artifactory.ap.org\n * updates: artifactory.ap.org\nNo package puppet-agent available.\n12:26:58 -0500 INFO: Return code: 1\n12:26:59 -0500 CRIT: Installation failed\n", "_error"=>{"kind"=>"puppetlabs.tasks/task-error", "issue_code"=>"TASK_ERROR", "msg"=>"The task failed with exit code 1", "details"=>{"exit_code"=>1}}}
  localhost:2226: {"_output"=>"12:26:47 -0500 INFO: Version parameter not defined and no agent detected. Assuming latest.\n12:26:47 -0500 INFO: Downloading Puppet latest for el...\n12:26:47 -0500 INFO: Red hat like platform! Lets get you an RPM...\n12:26:47 -0500 INFO: Downloading http://yum.puppet.com/puppet-release-el-7.noarch.rpm\n12:26:47 -0500 INFO:   to file /tmp/install.sh.51677.49907/puppet-release-el-7.noarch.rpm\n12:26:47 -0500 INFO: Trying curl...\n12:26:47 -0500 INFO: installing puppetlabs yum repo with rpm...\nLoaded plugins: fastestmirror, ovl\nLoading mirror speeds from cached hostfile\n * base: mirror.wdc1.us.leaseweb.net\n * epel: ewr.edge.kernel.org\n * extras: artifactory.ap.org\n * updates: packages.oit.ncsu.edu\nNo package puppet-agent available.\n12:26:49 -0500 INFO: Retrying... [1/5]\nLoaded plugins: fastestmirror, ovl\nLoading mirror speeds from cached hostfile\n * base: artifactory.ap.org\n * epel: ewr.edge.kernel.org\n * extras: artifactory.ap.org\n * updates: artifactory.ap.org\nNo package puppet-agent available.\n12:26:50 -0500 INFO: Return code: 1\n12:26:51 -0500 INFO: Retrying... [2/5]\nLoaded plugins: fastestmirror, ovl\nLoading mirror speeds from cached hostfile\n * base: mirror.wdc1.us.leaseweb.net\n * epel: ewr.edge.kernel.org\n * extras: artifactory.ap.org\n * updates: artifactory.ap.org\nNo package puppet-agent available.\n12:26:52 -0500 INFO: Return code: 1\n12:26:53 -0500 INFO: Retrying... [3/5]\nLoaded plugins: fastestmirror, ovl\nLoading mirror speeds from cached hostfile\n * base: mirror.wdc1.us.leaseweb.net\n * epel: iad.mirror.rackspace.com\n * extras: artifactory.ap.org\n * updates: packages.oit.ncsu.edu\nNo package puppet-agent available.\n12:26:54 -0500 INFO: Return code: 1\n12:26:55 -0500 INFO: Retrying... [4/5]\nLoaded plugins: fastestmirror, ovl\nLoading mirror speeds from cached hostfile\n * base: mirror.wdc1.us.leaseweb.net\n * epel: epel.mirror.constant.com\n * extras: artifactory.ap.org\n * updates: artifactory.ap.org\nNo package puppet-agent available.\n12:26:56 -0500 INFO: Return code: 1\n12:26:57 -0500 INFO: Retrying... [5/5]\nLoaded plugins: fastestmirror, ovl\nLoading mirror speeds from cached hostfile\n * base: mirror.wdc1.us.leaseweb.net\n * epel: mirror.math.princeton.edu\n * extras: mirror.metrocast.net\n * updates: packages.oit.ncsu.edu\nNo package puppet-agent available.\n12:26:58 -0500 INFO: Return code: 1\n12:26:59 -0500 CRIT: Installation failed\n", "_error"=>{"kind"=>"puppetlabs.tasks/task-error", "issue_code"=>"TASK_ERROR", "msg"=>"The task failed with exit code 1", "details"=>{"exit_code"=>1}}}}

Think it would also be useful for the target key instead of having to be localhost:222X it could be the list_name so you could do pdk bundle exec rake "litmus:tear_down[centos7-puppet6]" vs pdk bundle exec rake "litmus:tear_down[localhost:2222]"

[DavidS]

I like the idea in principle.

The problem I see is that there is no good connection from the provision_list invocation to agent_install task.

Using existing mechanisms in litmus, setting the desired puppet collection in the vars could work out. See

puppet_litmus/lib/puppet_litmus/rake_tasks.rb

Line 33 in 96510d3

inventory_vars = provision_hash[args[:key]]['vars']

vars are key/value pairs that get passed through provisioning and get stored in the inventory file, so the agent_install rake task would be able to pick this up again, and default :collection to that value.

Think it would also be useful for the target key instead of having to be localhost:222X it could be the list_name so you could do pdk bundle exec rake "litmus:tear_down[centos7-puppet6]" vs pdk bundle exec rake "litmus:tear_down[localhost:2222]"

It would be great to use more of bolt's target addressing features and map provision list entries to groups and allow users to set more useful aliases.

[cdenneen]

@DavidS do you have a working example of how vars could be picked up by agent_install task? Is this wired or would require some wiring. Sounds like a reasonable approach given how litmus is currently written to take advantage of rather than the original proposal.

As far as the tear_down or other tasks sometimes the localhost:222x approach is a bit clunky. Trying to figure out which host is which or even having a left over inventory file accidentally left around and then after provision the host is 2224 but during agent_install it fails because can't connect to 2222, 2223 from the inventory file.

[cdenneen]

@DavidS also is vars the params mentioned in the litmus documentation?

[cdenneen]

https://puppetlabs.github.io/litmus/Litmus-core-commands.html#provisioning-via-yaml

---
default:
  provisioner: docker
  images: ['litmusimage/centos:7']
centos7:
  provisioner: docker
  images: ['litmusimage/centos:7']
  params:
    param_a: someone
    param_b: something
  vars:
    var1=value1
    var2=value2

but the resulting inventory shows:

---
version: 2
groups:
- name: docker_nodes
  targets: []
- name: ssh_nodes
  targets:
  - uri: localhost:2227
    config:
      transport: ssh
      ssh:
        user: root
        password: root
        port: 2227
        host-key-check: false
    facts:
      provisioner: docker
      container_name: litmusimage_centos_7-2227
      platform: litmusimage/centos:7
      os-release:
        NAME: CentOS Linux
        VERSION: 7 (Core)
        ID: centos
        ID_LIKE: rhel fedora
        VERSION_ID: '7'
        PRETTY_NAME: CentOS Linux 7 (Core)
        ANSI_COLOR: 0;31
        CPE_NAME: cpe:/o:centos:centos:7
        HOME_URL: https://www.centos.org/
        BUG_REPORT_URL: https://bugs.centos.org/
        CENTOS_MANTISBT_PROJECT: CentOS-7
        CENTOS_MANTISBT_PROJECT_VERSION: '7'
        REDHAT_SUPPORT_PRODUCT: centos
        REDHAT_SUPPORT_PRODUCT_VERSION: '7'
    vars: var1=value1 var2=value2
- name: winrm_nodes
  targets: []

so based on that the docs need to be updated as params do not work?

[DavidS]

1. there's no explicit wiring around vars. litmus is all spit and bailing wire, all of the real work is done by bolt.
2. vars vs params: yeah that seems to be an oversight in the documentation.