search

saltstack / salt

Software to automate the management and configuration of any infrastructure or application at scale. Install Salt from the Salt package repositories here:
https://docs.saltproject.io/salt/install-guide/en/latest/
Apache License 2.0
14.19k stars 5.48k forks source link

salt-cloud delete with a map file fails when multiple providers defined #31579

Closed bradthurber closed 8 years ago

bradthurber commented 8 years ago

Description of Issue/Question

salt-cloud delete of a mapped environment fails when more than one provider is defined.

Example failing run: 

[root@s-cdc-jbsalt-01 cloud.maps.d]# salt-cloud -m /etc/salt/cloud.maps.d/map-uat2-jb -d
[INFO    ] salt-cloud starting
[INFO    ] Applying map from '/etc/salt/cloud.maps.d/map-uat2-jb'.
[ERROR   ] An un-handled exception was caught by salt's global exception handler:
KeyError: 'u-jbdom-201.adesa.com'
Traceback (most recent call last):
  File "/usr/bin/salt-cloud", line 10, in <module>
    salt_cloud()
  File "/usr/lib/python2.6/site-packages/salt/scripts.py", line 425, in salt_cloud
    client.run()
  File "/usr/lib/python2.6/site-packages/salt/cloud/cli.py", line 186, in run
    matching = mapper.delete_map(query='list_nodes')
  File "/usr/lib/python2.6/site-packages/salt/cloud/__init__.py", line 1673, in delete_map
    query_map = self.interpolated_map(query=query)
  File "/usr/lib/python2.6/site-packages/salt/cloud/__init__.py", line 1656, in interpolated_map
    names.remove(vm_name)
KeyError: 'u-jbdom-201.adesa.com'
Traceback (most recent call last):
  File "/usr/bin/salt-cloud", line 10, in <module>
    salt_cloud()
  File "/usr/lib/python2.6/site-packages/salt/scripts.py", line 425, in salt_cloud
    client.run()
  File "/usr/lib/python2.6/site-packages/salt/cloud/cli.py", line 186, in run
    matching = mapper.delete_map(query='list_nodes')
  File "/usr/lib/python2.6/site-packages/salt/cloud/__init__.py", line 1673, in delete_map
    query_map = self.interpolated_map(query=query)
  File "/usr/lib/python2.6/site-packages/salt/cloud/__init__.py", line 1656, in interpolated_map
    names.remove(vm_name)
KeyError: 'u-jbdom-201.adesa.com'

The workaround I have come up with is to hide all of the providers except for the one used by the map. The delete then works.

I'm wondering if there is perhaps some nuance with the cloud.providers.d file format that I am not seeing? All of my sanitized cloud files are below.

Setup

Add the following cloud configuration files

/etc/salt/cloud.maps.d/map-uat2-jb

prof-vmware-nonprod-jbdom:
  - u-jbdom-201.adesa.com:
      grains:
        environment: uat2

/etc/salt/cloud.profiles.d/prof-vmware-nonprod-jb.conf

prof-vmware-nonprod-jb-instances:
  provider: prov-vmware-nonprod-jb

  clonefrom: rhel6xnonprodtemplate.company.com

  datacenter: CAR
  datastore: Pure-CDC
  cluster: CVEP-LX
  folder: JBOSS

  ssh_username: Brad.Thurber
  password: SANITIZED
  deploy: True
  domain: company.com

prof-vmware-nonprod-jbdom:
  cores_per_socket: 2
  devices:
    network:
      Network adapter 1:
        name: VLAN641
        switch_type: standard
  extends: prof-vmware-nonprod-jb-instances
  grains:
    jboss_domain_controller: True
    platform:
      name: jboss
      ou: adesac
      realm: nonprod 
    roles:
      - jboss
  memory: 8GB
  num_cpus: 2

prof-vmware-nonprod-jbapp:
  cores_per_socket: 2
  devices:
    network:
      Network adapter 1:
        name: VLAN641
        switch_type: standard
  extends: prof-vmware-nonprod-jb-instances
  grains:
    platform:
      name: jboss
      ou: adesac
      realm: nonprod
    roles:
      - jboss
      - mod-cluster-node
  memory: 64GB
  num_cpus: 8

prof-vmware-nonprod-jbweb:
  devices:
    network:
      Network adapter 1:
        name: VLAN230
        switch_type: standard
  extends: prof-vmware-nonprod-jb-instances
  grains:
    platform:
      name: jboss
      ou: adesac
      realm: nonprod
    roles:
      - webserver
      - mod-cluster-server
  memory: 4GB
  num_cpus: 1

/etc/salt/cloud.profiles.d/prof-vmware-prod-jb.conf

prof-vmware-prod-jb-instances:
  provider: prov-vmware-prod-jb

  clonefrom: rhel6xnonprodtemplate.company.com

  datacenter: LIFELINE
  datastore: Pure-CDC
  cluster: PVEP-LX
  folder: JBOSS

  ssh_username: Brad.Thurber
  password: SANITIZED
  deploy: True
  domain: company.com

prof-vmware-prod-jbdom:
  extends: prof-vmware-prod-jb-instances
  cores_per_socket: 2
  devices:
    network:
      Network adapter 1:
        name: VLAN170
        switch_type: standard
  grains:
    jboss_domain_controller: True
    roles:
      - jboss
  memory: 8GB
  num_cpus: 2

prof-vmware-prod-jbapp:
  extends: prof-vmware-prod-jb-instances
  cores_per_socket: 2
  devices:
    network:
      Network adapter 1:
        name: VLAN170
        switch_type: standard
  grains:
    roles:
      - jboss
      - mod-cluster-node
  memory: 64GB
  num_cpus: 8

prof-vmware-prod-jbweb:
  extends: prof-vmware-prod-jb-instances
  devices:
    network:
      Network adapter 1:
        name: VLAN765
        switch_type: standard
  grains:
    roles:
      - webserver
      - mod-cluster-server
  memory: 4GB
  num_cpus: 1

/etc/salt/cloud.providers.d/prov_vmware_nonprod.conf

prov-vmware-nonprod-jb:
  driver: vmware
  user: 'DOMAIN\username'
  password: 'SANITIZED'
  url: 'p-corp-vcntr.company.com'
  protocol: 'https'
  port: 443
  minion:
    master: saltmaster.company.com

/etc/salt/cloud.providers.d/prov_vmware_prod.conf

prov-vmware-prod-jb:
  driver: vmware
  user: 'DOMAIN\username'
  password: 'SANITIZED'
  url: 'p-corp-vcntr.company.com'
  protocol: 'https'
  port: 443
  minion:
    master: saltmaster.company.com

Steps to Reproduce Issue

Run salt-cloud map delete as shown in the description above. You should get the failure as in the description above.

Now hide the prov_vmware_prod.conf file (this is the one that is NOT used by the map) by renaming it to prov_vmware_prod.confHIDDEN. The salt cloud map delete operation will now work.

Versions Report

Provided by running salt --versions-report

[root@s-cdc-jbsalt-01 salt]# salt --versions-report                                                                                                     [2/9611]Salt Version:
           Salt: 2015.8.7

Dependency Versions:
         Jinja2: 2.8
       M2Crypto: 0.20.2
           Mako: 0.3.4
         PyYAML: 3.11
          PyZMQ: 14.5.0
         Python: 2.6.6 (r266:84292, Nov 21 2013, 10:50:32)
           RAET: Not Installed
        Tornado: 4.2.1
            ZMQ: 4.0.5
           cffi: Not Installed
       cherrypy: 3.2.2
       dateutil: 1.4.1
          gitdb: 0.5.4
      gitpython: 0.3.2 RC1
          ioflo: Not Installed
        libgit2: 0.20.0
        libnacl: Not Installed
   msgpack-pure: Not Installed
 msgpack-python: 0.4.6
   mysql-python: Not Installed
      pycparser: Not Installed
       pycrypto: 2.6.1
         pygit2: 0.20.3
   python-gnupg: Not Installed
          smmap: 0.8.1
        timelib: Not Installed

System Versions:
           dist: redhat 6.6 Santiago
        machine: x86_64
        release: 2.6.32-504.el6.x86_64
         system: Red Hat Enterprise Linux Server 6.6 Santiago
jfindlay commented 8 years ago

@bradthurber, thanks for reporting.

rickh563 commented 8 years ago

Salt-Support has validated that this is a bug, though with the following caveats:

This bug only shows up when using vmware with salt-cloud and using the extends profile functionality.

  1. The bug has been confirmed with the vmware driver.
  2. The bug was not confirmed with our test case using amazon or ec2 driver.
  3. Salt-Support confirmed that the workaround works just as the customer stated.
  4. The bug does not show up when we remove the extends from the profile configurations.

For additional detail see:

rickh563 commented 8 years ago

ZD-665

rallytime commented 8 years ago

@bradthurber Does the fix in #31719 resolve this issue for you? That fix is included in the 2015.8.8 release.

bradthurber commented 8 years ago

@rallytime, yes, this is fixed now. Closing

rallytime commented 8 years ago

Awesome! Thanks!