Uploaded image for project: 'Puppet Agent'
  1. Puppet Agent
  2. PA-67

Mcollective subsystem never leaves inoperative state in AIX

    Details

    • Type: Bug
    • Status: Closed
    • Priority: Normal
    • Resolution: Fixed
    • Affects Version/s: puppet-agent 1.2.7
    • Fix Version/s: puppet-agent 1.3.0
    • Component/s: None
    • Labels:
      None
    • Environment:

      AIX 5.3, 6.1 or 7.1.

    • Template:
    • Story Points:
      2
    • Sprint:
      Client 2015-11-11
    • Release Notes:
      Bug Fix
    • Release Notes Summary:
      Hide
      Previously in puppet-agent on AIX, the mcollective service was improperly daemonized when starting the service, causing the AIX subsystem manager to lose track of the child (daemon) process. This meant that the service always appeared to be 'inoperative', even if it was really running. The mcollective subsystem command arguments have been updated to include the '--no-daemonize' option to allow the subsystem manager to properly handle the service.
      Show
      Previously in puppet-agent on AIX, the mcollective service was improperly daemonized when starting the service, causing the AIX subsystem manager to lose track of the child (daemon) process. This meant that the service always appeared to be 'inoperative', even if it was really running. The mcollective subsystem command arguments have been updated to include the '--no-daemonize' option to allow the subsystem manager to properly handle the service.

      Description

      In AIX, when starting the mcollective service, we see:

      [0] [AIX] root@pe-aix-61-agent:~/new # startsrc -s mcollective
      0513-059 The mcollective Subsystem has been started. Subsystem PID is 9568436.
      

      ... which would seem to indicate that the subsystem was successfully started with PID 9568436. However, when checking on the status of the service, it always comes back as inoperative:

      [0] [AIX] root@pe-aix-61-agent:~/new # lssrc -s mcollective
      Subsystem         Group            PID          Status
       mcollective                                    inoperative
      

      Searching for the mcollective process, we get:

      [0] [AIX] root@pe-aix-61-agent:~/new # ps -elf | grep mcollective
         40001 A     root 10289346        1   0  60 20 83c679480  9516        * 11:08:13      -  0:00 /opt/puppetlabs/puppet/bin/ruby /opt/puppetlabs/puppet/bin/mcollectived --config=/etc/puppetlabs/mcollective/server.cfg
      

      ... which is not the same PID as `startsrc` reported. In addition, in /var/log/puppetlabs/mcollective.log, we see many (expected) activemq connection failures:

      [0] [AIX] root@pe-aix-61-agent:~/new # tail -f /var/log/puppetlabs/mcollective.log
      I, [2015-11-09T11:09:36.220912 #10813554]  INFO -- : activemq.rb:114:in `on_connecting' TCP Connection attempt 5 to stomp://mcollective@stomp1:6163
      I, [2015-11-09T11:09:36.224217 #10813554]  INFO -- : activemq.rb:129:in `on_connectfail' TCP Connection to stomp://mcollective@stomp1:6163 failed on attempt 5
      

      This is expected, and is the same as what we see on other platforms when running mcollective without properly configuring it. Given all of this, it seems that the subsystem should be reported as 'running' in this state (like other platforms), but something is preventing the subsystem to leave the 'inoperative' state.

        Attachments

          Issue Links

            Activity

              jsd-sla-details-panel

                People

                • Assignee:
                  john.duarte John Duarte
                  Reporter:
                  whopper William Hopper
                • Votes:
                  0 Vote for this issue
                  Watchers:
                  3 Start watching this issue

                  Dates

                  • Created:
                    Updated:
                    Resolved:

                    Zendesk Support