Class: PuppetX::Relay::Agent::Job::Exec

Inherits:
Base
  • Object
show all
Defined in:
lib/puppet_x/relay/agent/job/exec.rb

Instance Method Summary collapse

Methods inherited from Base

#to_job

Constructor Details

#initialize(backend, run, state_dir) ⇒ Exec

Returns a new instance of Exec.

Parameters:



17
18
19
20
21
22
23
 
# File 'lib/puppet_x/relay/agent/job/exec.rb', line 17

def initialize(backend, run, state_dir)
  @backend = backend
  @run = run
  @state_dir = state_dir
  @accepted = false
  @retries = 3
end

Instance Method Details

#handle(job) ⇒ Object 



25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
 
# File 'lib/puppet_x/relay/agent/job/exec.rb', line 25

def handle(job)
  Puppet::FileSystem.mkpath(@state_dir)

  @state_lock ||= Puppet::Util::Pidlock.new(File.join(@state_dir, 'state'))
  unless @state_lock.lock
    # Running concurrently with another process.
    job.disable
    return
  end

  begin
    sched = Schedule.from_scheduler_job(job)
    stamp_file = File.join(@state_dir, 'stamp')

    # If there's a stamp file, we know that this node should be
    # working on this run. If the stamp is newer than our run, another
    # job on this same node processed it first.
    begin
      @accepted ||= Puppet::FileSystem.stat(stamp_file).mtime <= @run.state.updated_at
    rescue Errno::ENOENT # rubocop:disable Lint/HandleExceptions
    end

    # NB: There is a (minor) race condition here where the process
    # could exit (SIGKILL or hard crash) between when the run is
    # accepted by the API and when the acceptance state file is
    # written out. We should update the API to take a unique token
    # that we generate to make acceptance idempotent.
    @accepted ||= @backend.relay_api.post_accept_run(@run)
    if @accepted
      Puppet::FileSystem.touch(stamp_file, mtime: @run.state.updated_at)
    else
      Puppet.notice(_('Run %{id} is already being processed by another job, ignoring') % { id: @run.id })
      job.disable
      return
    end

    if @run.state.status == :pending
      Puppet.notice(_('Run %{id} started') % { id: @run.id })
    end

    begin
      @run = @backend.exec(@run, @state_dir, sched)
    rescue NotImplementedError => e
      Puppet.log_exception(e, _('The backend %{backend_name} does not support the configuration for run %{id}: %{message}') % {
        backend_name: @backend.class.name,
        id: @run.id,
        message: e.message,
      })
      @run = @run.with_state(@run.state.to_complete(outcome: 'error'))
    rescue StandardError => e
      Puppet.log_exception(e, _('Run %{id} encountered an error during execution: %{message} (%{retries} retries remaining)') % { id: @run.id, message: e.message, retries: @retries })

      if (@retries -= 1) < 0
        Puppet.warning(_('Retries exhausted for run %{id}, transitioning to complete with error outcome') % { id: @run.id })
        @run = @run.with_state(@run.state.to_complete(outcome: 'error'))
      end
    ensure
      @run = @backend.relay_api.put_run_state(@run)
      Puppet::FileSystem.touch(stamp_file, mtime: @run.state.updated_at)
    end

    if @run.state.status == :complete
      Puppet.notice(_('Run %{id} completed with outcome %{outcome}') % { id: @run.id, outcome: @run.state.outcome || _('(unknown)') })
      job.disable
    end
  ensure
    @state_lock.unlock

    # If we're finished, we can remove the state directory entirely.
    FileUtils.remove_entry_secure(@state_dir, true) unless job.enabled?
  end
end