Pillar unavailable in orchestrate runner

[rbjorklin]

Description of Issue/Question

It is not possible to use pillar values from /srv/pillar with the orchestrate runner. However the pillar values passed in on command line as in the below example are available. Pillar values provided by the ext_pillar consul are also available. Is there some kind of (un?)documented way of distributing pillars to the master?

Setup

Run with: salt-run --out=json state.orchestrate orch.application.upgrade pillar='{"data":{"organization":"company","cluster":"demo","version":"1.0"}}'

# in /srv/pillar/application/managers.sls
managers:
  company:
    hostname: demo.example.org

# in /srv/salt/orch/application/upgrade.sls
upgrade-application-manager-{{ pillar['data']['organization'] }}-{{ pillar['data']['cluster'] }}:
  salt.state:
    - tgt: {{ pillar['managers'][pillar['data']['organization']]['hostname'] }}
    - sls:
      - test.ping
    - failhard: True

Running the above results in:

[ERROR   ] Rendering exception occurred: Jinja variable 'salt.utils.context.NamespacedDictWrapper object' has no attribute 'managers'
[CRITICAL] Rendering SLS 'base:orch.application.upgrade' failed: Jinja variable 'salt.utils.context.NamespacedDictWrapper object' has no attribute 'managers'
{
    "retcode": 1,
    "salt-master.localdomain_master": [
        "Rendering SLS 'base:orch.application.upgrade' failed: Jinja variable 'salt.utils.context.NamespacedDictWrapper object' has no attribute 'managers'"
    ]
}

Steps to Reproduce Issue

(Include debug logs if possible and relevant.)

Versions Report

Salt Version:
           Salt: 2016.3.0

Dependency Versions:
           cffi: 0.8.6
       cherrypy: 3.2.2
       dateutil: 1.5
          gitdb: Not Installed
      gitpython: Not Installed
          ioflo: Not Installed
         Jinja2: 2.7.2
        libgit2: Not Installed
        libnacl: Not Installed
       M2Crypto: 0.21.1
           Mako: Not Installed
   msgpack-pure: Not Installed
 msgpack-python: 0.4.7
   mysql-python: Not Installed
      pycparser: 2.14
       pycrypto: 2.6.1
         pygit2: Not Installed
         Python: 2.7.5 (default, Nov 20 2015, 02:00:19)
   python-gnupg: Not Installed
         PyYAML: 3.11
          PyZMQ: 14.7.0
           RAET: Not Installed
          smmap: Not Installed
        timelib: Not Installed
        Tornado: 4.2.1
            ZMQ: 4.0.5

System Versions:
           dist: centos 7.2.1511 Core
        machine: x86_64
        release: 3.10.0-327.18.2.el7.x86_64
         system: Linux
        version: CentOS Linux 7.2.1511 Core

[rbjorklin]

This seems to be heavily related to #9442. Is there any way to call a runner from within a orchestrate state? I.E.

{% set master_pillar = salt-run['pillar.show_pillar']('salt.example.org') %}

See https://docs.saltstack.com/en/latest/ref/runners/all/salt.runners.pillar.html

[Ch3LL]

@rbjorklin I think there are two issues going on here.

First {{ pillar['managers'][pillar['data']['organization']]['hostname'] }} will cause some issues. I'm guessing what you want is this: {{ pillar['managers']['company']['hostname' }} maybe?If you want to call two pillars you would have to seperate them out.

Also I think you might be running into this issue: #33647 which caused issues with pillars on the command line overriding pillar in /srv/pillar. This has been fixed in 2016.3.1. Once you fix that pillar is it working ?

[rbjorklin]

@Ch3LL I use nested pillar calls in several other states so unless there is something different with orchestration that should be fine I think.

It does indeed look like I'm running into #33647 however upgrading to 2016.3.1 seems to make me run into #29028 which is actually a lot worse in my case.

EDIT: Just upgraded to 2016.3.1 and it seems to work, so far #29028 hasn't resurfaced.

[rbjorklin]

Looks like #33647 was my issue. Distributing pillars to the master is done by appending _master to the minion name of the minion running on the master (if you have one). Closing this issue.