Merge pull request #6541 from smarterclayton/move_gcp_in

Allow openshift-ansible image to deploy to GCP

7 files changed, 515 insertions, 0 deletions

diff --git a/inventory/dynamic/gcp/README.md b/inventory/dynamic/gcp/README.md
new file mode 100644
index 000000000..217a035ca
--- /dev/null
+++ b/inventory/dynamic/gcp/README.md
@@ -0,0 +1 @@
+This directory provides dynamic inventory for a GCP cluster configured via the GCP provisioning playbook. Set inventory to `inventory/dynamic/gcp/hosts.sh` to calculate the appropriate host set.
\ No newline at end of file
diff --git a/inventory/dynamic/gcp/ansible.cfg b/inventory/dynamic/gcp/ansible.cfg
new file mode 100644
index 000000000..f87d51f28
--- /dev/null
+++ b/inventory/dynamic/gcp/ansible.cfg
@@ -0,0 +1,45 @@
+# config file for ansible -- http://ansible.com/
+# ==============================================
+
+# This config file provides examples for running
+# the OpenShift playbooks with the provided
+# inventory scripts.
+
+[defaults]
+# Set the log_path
+#log_path = /tmp/ansible.log
+
+private_key_file = $HOME/.ssh/google_compute_engine
+
+# Additional default options for OpenShift Ansible
+forks = 50
+host_key_checking = False
+retry_files_enabled = False
+retry_files_save_path = ~/ansible-installer-retries
+nocows = True
+remote_user = cloud-user
+roles_path = ../../../roles/
+gathering = smart
+fact_caching = jsonfile
+fact_caching_connection = $HOME/ansible/facts
+fact_caching_timeout = 600
+callback_whitelist = profile_tasks
+inventory_ignore_extensions = secrets.py, .pyc, .cfg, .crt
+# work around privilege escalation timeouts in ansible:
+timeout = 30
+
+# Uncomment to use the provided example inventory
+inventory = hosts.sh
+
+[inventory]
+# fail more helpfully when the inventory file does not parse (Ansible 2.4+)
+unparsed_is_failed=true
+
+# Additional ssh options for OpenShift Ansible
+[ssh_connection]
+pipelining = True
+ssh_args = -o ControlMaster=auto -o ControlPersist=600s
+timeout = 10
+# shorten the ControlPath which is often too long; when it is,
+# ssh connection reuse silently fails, making everything slower.
+control_path = %(directory)s/%%h-%%r
diff --git a/inventory/dynamic/gcp/group_vars/all/00_defaults.yml b/inventory/dynamic/gcp/group_vars/all/00_defaults.yml
new file mode 100644
index 000000000..2f72e905f
--- /dev/null
+++ b/inventory/dynamic/gcp/group_vars/all/00_defaults.yml
@@ -0,0 +1,42 @@
+# GCP uses non-root users by default, so sudo by default
+---
+ansible_become: yes
+
+openshift_deployment_type: origin
+
+# Debugging settings
+debug_level: 2
+openshift_debug_level: "{{ debug_level }}"
+openshift_master_debug_level: "{{ master_debug_level | default(debug_level, true) }}"
+openshift_node_debug_level: "{{ node_debug_level | default(debug_level, true) }}"
+
+# External API settings
+console_port: 443
+internal_console_port: 8443
+openshift_master_api_port: "8443"
+openshift_master_console_port: "8443"
+openshift_master_cluster_hostname: "internal-openshift-master.{{ public_hosted_zone }}"
+openshift_master_cluster_public_hostname: "openshift-master.{{ public_hosted_zone }}"
+openshift_master_default_subdomain: "{{ wildcard_zone }}"
+
+# Cloud specific settings
+openshift_cloudprovider_kind: gce
+openshift_hosted_registry_storage_provider: gcs
+
+openshift_master_access_token_max_seconds: 2419200
+openshift_master_identity_providers:
+
+# Networking settings
+openshift_node_port_range: 30000-32000
+openshift_node_open_ports: [{"service":"Router stats port", "port":"1936/tcp"}, {"service":"Allowed open host ports", "port":"9000-10000/tcp"}, {"service":"Allowed open host ports", "port":"9000-10000/udp"}]
+openshift_node_sdn_mtu: 1410
+osm_cluster_network_cidr: 172.16.0.0/16
+osm_host_subnet_length: 9
+openshift_portal_net: 172.30.0.0/16
+
+# Default cluster configuration
+openshift_master_cluster_method: native
+openshift_schedulable: true
+# TODO: change to upstream conventions
+openshift_hosted_infra_selector: "role=infra"
+osm_default_node_selector: "role=app"
diff --git a/inventory/dynamic/gcp/hosts.py b/inventory/dynamic/gcp/hosts.py
new file mode 100755
index 000000000..cd1262622
--- /dev/null
+++ b/inventory/dynamic/gcp/hosts.py
@@ -0,0 +1,408 @@
+#!/usr/bin/env python
+# Copyright 2013 Google Inc.
+#
+# This file is part of Ansible
+#
+# Ansible is free software: you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation, either version 3 of the License, or
+# (at your option) any later version.
+#
+# Ansible is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with Ansible.  If not, see <http://www.gnu.org/licenses/>.
+
+# This is a derivative of gce.py that adds support for filtering
+# the returned inventory to only include instances that have tags
+# as specified by GCE_TAGGED_INSTANCES. This prevents dynamic 
+# inventory for multiple clusters within the same project from
+# accidentally stomping each other.
+
+# pylint: skip-file
+
+'''
+GCE external inventory script
+=================================
+
+Generates inventory that Ansible can understand by making API requests
+Google Compute Engine via the libcloud library.  Full install/configuration
+instructions for the gce* modules can be found in the comments of
+ansible/test/gce_tests.py.
+
+When run against a specific host, this script returns the following variables
+based on the data obtained from the libcloud Node object:
+ - gce_uuid
+ - gce_id
+ - gce_image
+ - gce_machine_type
+ - gce_private_ip
+ - gce_public_ip
+ - gce_name
+ - gce_description
+ - gce_status
+ - gce_zone
+ - gce_tags
+ - gce_metadata
+ - gce_network
+
+When run in --list mode, instances are grouped by the following categories:
+ - zone:
+   zone group name examples are us-central1-b, europe-west1-a, etc.
+ - instance tags:
+   An entry is created for each tag.  For example, if you have two instances
+   with a common tag called 'foo', they will both be grouped together under
+   the 'tag_foo' name.
+ - network name:
+   the name of the network is appended to 'network_' (e.g. the 'default'
+   network will result in a group named 'network_default')
+ - machine type
+   types follow a pattern like n1-standard-4, g1-small, etc.
+ - running status:
+   group name prefixed with 'status_' (e.g. status_running, status_stopped,..)
+ - image:
+   when using an ephemeral/scratch disk, this will be set to the image name
+   used when creating the instance (e.g. debian-7-wheezy-v20130816).  when
+   your instance was created with a root persistent disk it will be set to
+   'persistent_disk' since there is no current way to determine the image.
+
+Examples:
+  Execute uname on all instances in the us-central1-a zone
+  $ ansible -i gce.py us-central1-a -m shell -a "/bin/uname -a"
+
+  Use the GCE inventory script to print out instance specific information
+  $ contrib/inventory/gce.py --host my_instance
+
+Author: Eric Johnson <erjohnso@google.com>
+Contributors: Matt Hite <mhite@hotmail.com>
+Version: 0.0.2
+'''
+
+__requires__ = ['pycrypto>=2.6']
+try:
+    import pkg_resources
+except ImportError:
+    # Use pkg_resources to find the correct versions of libraries and set
+    # sys.path appropriately when there are multiversion installs.  We don't
+    # fail here as there is code that better expresses the errors where the
+    # library is used.
+    pass
+
+USER_AGENT_PRODUCT="Ansible-gce_inventory_plugin"
+USER_AGENT_VERSION="v2"
+
+import sys
+import os
+import time
+import argparse
+import ConfigParser
+
+import logging
+logging.getLogger('libcloud.common.google').addHandler(logging.NullHandler())
+
+try:
+    import json
+except ImportError:
+    import simplejson as json
+
+try:
+    from libcloud.compute.types import Provider
+    from libcloud.compute.providers import get_driver
+    from libcloud.common.google import ResourceNotFoundError
+    _ = Provider.GCE
+except:
+    sys.exit("GCE inventory script requires libcloud >= 0.13")
+
+
+class GceInventory(object):
+    def __init__(self):
+        # Read settings and parse CLI arguments
+        self.parse_cli_args()
+        self.config = self.get_config()
+        self.driver = self.get_gce_driver()
+        self.ip_type = self.get_inventory_options()
+        if self.ip_type:
+            self.ip_type = self.ip_type.lower()
+
+        # Just display data for specific host
+        if self.args.host:
+            print(self.json_format_dict(self.node_to_dict(
+                    self.get_instance(self.args.host)),
+                    pretty=self.args.pretty))
+            sys.exit(0)
+
+        zones = self.parse_env_zones()
+
+        # Otherwise, assume user wants all instances grouped
+        print(self.json_format_dict(self.group_instances(zones),
+            pretty=self.args.pretty))
+        sys.exit(0)
+
+    def get_config(self):
+        """
+        Populates a SafeConfigParser object with defaults and
+        attempts to read an .ini-style configuration from the filename
+        specified in GCE_INI_PATH. If the environment variable is
+        not present, the filename defaults to gce.ini in the current
+        working directory.
+        """
+        gce_ini_default_path = os.path.join(
+            os.path.dirname(os.path.realpath(__file__)), "gce.ini")
+        gce_ini_path = os.environ.get('GCE_INI_PATH', gce_ini_default_path)
+
+        # Create a ConfigParser.
+        # This provides empty defaults to each key, so that environment
+        # variable configuration (as opposed to INI configuration) is able
+        # to work.
+        config = ConfigParser.SafeConfigParser(defaults={
+            'gce_service_account_email_address': '',
+            'gce_service_account_pem_file_path': '',
+            'gce_project_id': '',
+            'libcloud_secrets': '',
+            'inventory_ip_type': '',
+        })
+        if 'gce' not in config.sections():
+            config.add_section('gce')
+        if 'inventory' not in config.sections():
+            config.add_section('inventory')
+
+        config.read(gce_ini_path)
+
+        #########
+        # Section added for processing ini settings
+        #########
+
+        # Set the instance_states filter based on config file options
+        self.instance_states = []
+        if config.has_option('gce', 'instance_states'):
+            states = config.get('gce', 'instance_states')
+            # Ignore if instance_states is an empty string.
+            if states:
+                self.instance_states = states.split(',')
+
+        return config
+
+    def get_inventory_options(self):
+        """Determine inventory options. Environment variables always
+        take precedence over configuration files."""
+        ip_type = self.config.get('inventory', 'inventory_ip_type')
+        # If the appropriate environment variables are set, they override
+        # other configuration
+        ip_type = os.environ.get('INVENTORY_IP_TYPE', ip_type)
+        return ip_type
+
+    def get_gce_driver(self):
+        """Determine the GCE authorization settings and return a
+        libcloud driver.
+        """
+        # Attempt to get GCE params from a configuration file, if one
+        # exists.
+        secrets_path = self.config.get('gce', 'libcloud_secrets')
+        secrets_found = False
+        try:
+            import secrets
+            args = list(getattr(secrets, 'GCE_PARAMS', []))
+            kwargs = getattr(secrets, 'GCE_KEYWORD_PARAMS', {})
+            secrets_found = True
+        except:
+            pass
+
+        if not secrets_found and secrets_path:
+            if not secrets_path.endswith('secrets.py'):
+                err = "Must specify libcloud secrets file as "
+                err += "/absolute/path/to/secrets.py"
+                sys.exit(err)
+            sys.path.append(os.path.dirname(secrets_path))
+            try:
+                import secrets
+                args = list(getattr(secrets, 'GCE_PARAMS', []))
+                kwargs = getattr(secrets, 'GCE_KEYWORD_PARAMS', {})
+                secrets_found = True
+            except:
+                pass
+        if not secrets_found:
+            args = [
+                self.config.get('gce','gce_service_account_email_address'),
+                self.config.get('gce','gce_service_account_pem_file_path')
+            ]
+            kwargs = {'project': self.config.get('gce', 'gce_project_id')}
+
+        # If the appropriate environment variables are set, they override
+        # other configuration; process those into our args and kwargs.
+        args[0] = os.environ.get('GCE_EMAIL', args[0])
+        args[1] = os.environ.get('GCE_PEM_FILE_PATH', args[1])
+        kwargs['project'] = os.environ.get('GCE_PROJECT', kwargs['project'])
+
+        # Retrieve and return the GCE driver.
+        gce = get_driver(Provider.GCE)(*args, **kwargs)
+        gce.connection.user_agent_append(
+            '%s/%s' % (USER_AGENT_PRODUCT, USER_AGENT_VERSION),
+        )
+        return gce
+
+    def parse_env_zones(self):
+        '''returns a list of comma seperated zones parsed from the GCE_ZONE environment variable.
+        If provided, this will be used to filter the results of the grouped_instances call'''
+        import csv
+        reader = csv.reader([os.environ.get('GCE_ZONE',"")], skipinitialspace=True)
+        zones = [r for r in reader]
+        return [z for z in zones[0]]
+
+    def parse_cli_args(self):
+        ''' Command line argument processing '''
+
+        parser = argparse.ArgumentParser(
+                description='Produce an Ansible Inventory file based on GCE')
+        parser.add_argument('--list', action='store_true', default=True,
+                           help='List instances (default: True)')
+        parser.add_argument('--host', action='store',
+                           help='Get all information about an instance')
+        parser.add_argument('--tagged', action='store',
+                           help='Only include instances with this tag')
+        parser.add_argument('--pretty', action='store_true', default=False,
+                           help='Pretty format (default: False)')
+        self.args = parser.parse_args()
+
+        tag_env = os.environ.get('GCE_TAGGED_INSTANCES')
+        if not self.args.tagged and tag_env:
+            self.args.tagged = tag_env
+
+    def node_to_dict(self, inst):
+        md = {}
+
+        if inst is None:
+            return {}
+
+        if inst.extra['metadata'].has_key('items'):
+            for entry in inst.extra['metadata']['items']:
+                md[entry['key']] = entry['value']
+
+        net = inst.extra['networkInterfaces'][0]['network'].split('/')[-1]
+        # default to exernal IP unless user has specified they prefer internal
+        if self.ip_type == 'internal':
+            ssh_host = inst.private_ips[0]
+        else:
+            ssh_host = inst.public_ips[0] if len(inst.public_ips) >= 1 else inst.private_ips[0]
+
+        return {
+            'gce_uuid': inst.uuid,
+            'gce_id': inst.id,
+            'gce_image': inst.image,
+            'gce_machine_type': inst.size,
+            'gce_private_ip': inst.private_ips[0],
+            'gce_public_ip': inst.public_ips[0] if len(inst.public_ips) >= 1 else None,
+            'gce_name': inst.name,
+            'gce_description': inst.extra['description'],
+            'gce_status': inst.extra['status'],
+            'gce_zone': inst.extra['zone'].name,
+            'gce_tags': inst.extra['tags'],
+            'gce_metadata': md,
+            'gce_network': net,
+            # Hosts don't have a public name, so we add an IP
+            'ansible_host': ssh_host
+        }
+
+    def get_instance(self, instance_name):
+        '''Gets details about a specific instance '''
+        try:
+            return self.driver.ex_get_node(instance_name)
+        except Exception as e:
+            return None
+
+    def group_instances(self, zones=None):
+        '''Group all instances'''
+        groups = {}
+        meta = {}
+        meta["hostvars"] = {}
+
+        # list_nodes will fail if a disk is in the process of being deleted
+        # from a node, which is not uncommon if other playbooks are managing
+        # the same project. Retry if we receive a not found error.
+        nodes = []
+        tries = 0
+        while True:
+            try:
+                nodes = self.driver.list_nodes()
+                break
+            except ResourceNotFoundError:
+                tries = tries + 1
+                if tries > 15:
+                    raise e
+                time.sleep(1)
+                continue
+
+        for node in nodes:
+
+            # This check filters on the desired instance states defined in the
+            # config file with the instance_states config option.
+            #
+            # If the instance_states list is _empty_ then _ALL_ states are returned.
+            #
+            # If the instance_states list is _populated_ then check the current
+            # state against the instance_states list
+            if self.instance_states and not node.extra['status'] in self.instance_states:
+                continue
+
+            name = node.name
+
+            if self.args.tagged and self.args.tagged not in node.extra['tags']:
+                continue
+
+            meta["hostvars"][name] = self.node_to_dict(node)
+
+            zone = node.extra['zone'].name
+
+            # To avoid making multiple requests per zone
+            # we list all nodes and then filter the results
+            if zones and zone not in zones:
+                continue
+
+            if groups.has_key(zone): groups[zone].append(name)
+            else: groups[zone] = [name]
+
+            tags = node.extra['tags']
+            for t in tags:
+                if t.startswith('group-'):
+                    tag = t[6:]
+                else:
+                    tag = 'tag_%s' % t
+                if groups.has_key(tag): groups[tag].append(name)
+                else: groups[tag] = [name]
+
+            net = node.extra['networkInterfaces'][0]['network'].split('/')[-1]
+            net = 'network_%s' % net
+            if groups.has_key(net): groups[net].append(name)
+            else: groups[net] = [name]
+
+            machine_type = node.size
+            if groups.has_key(machine_type): groups[machine_type].append(name)
+            else: groups[machine_type] = [name]
+
+            image = node.image and node.image or 'persistent_disk'
+            if groups.has_key(image): groups[image].append(name)
+            else: groups[image] = [name]
+
+            status = node.extra['status']
+            stat = 'status_%s' % status.lower()
+            if groups.has_key(stat): groups[stat].append(name)
+            else: groups[stat] = [name]
+
+        groups["_meta"] = meta
+
+        return groups
+
+    def json_format_dict(self, data, pretty=False):
+        ''' Converts a dict to a JSON object and dumps it as a formatted
+        string '''
+
+        if pretty:
+            return json.dumps(data, sort_keys=True, indent=2)
+        else:
+            return json.dumps(data)
+
+
+# Run the script
+GceInventory()
diff --git a/inventory/dynamic/gcp/hosts.sh b/inventory/dynamic/gcp/hosts.sh
new file mode 100755
index 000000000..0c88e3a6b
--- /dev/null
+++ b/inventory/dynamic/gcp/hosts.sh
@@ -0,0 +1,15 @@
+#!/bin/sh
+
+set -euo pipefail
+
+# Use a playbook to calculate the inventory dynamically from
+# the provided cluster variables.
+src="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+if ! out="$( ansible-playbook --inventory-file "${src}/none" ${src}/../../../playbooks/gcp/openshift-cluster/inventory.yml 2>&1 )"; then
+  echo "error: Inventory configuration failed" 1>&2
+  echo "$out" 1>&2
+  echo "{}"
+  exit 1
+fi
+source "/tmp/inventory.sh"
+exec ${src}/hosts.py
diff --git a/inventory/dynamic/gcp/none b/inventory/dynamic/gcp/none
new file mode 100644
index 000000000..9e26dfeeb
--- /dev/null
+++ b/inventory/dynamic/gcp/none
@@ -0,0 +1 @@
+{}
\ No newline at end of file
diff --git a/inventory/dynamic/injected/README.md b/inventory/dynamic/injected/README.md
new file mode 100644
index 000000000..5e2e4c549
--- /dev/null
+++ b/inventory/dynamic/injected/README.md
@@ -0,0 +1,3 @@
+This directory may be used to inject inventory into openshift-ansible
+when used in a container. Other scripts like the cloud provider entrypoints
+will automatically use the content of this directory as inventory.