From 6f61581dcc89b33fec2fe3916b0d512378e370ca Mon Sep 17 00:00:00 2001 From: Thomas Wiest Date: Fri, 21 Aug 2015 16:01:33 -0400 Subject: updating to the latest ec2.py (and re-patching with our changes). --- inventory/aws/hosts/ec2.ini | 97 ++++++- inventory/aws/hosts/ec2.py | 645 +++++++++++++++++++++++++++++++++++++++----- 2 files changed, 680 insertions(+), 62 deletions(-) (limited to 'inventory') diff --git a/inventory/aws/hosts/ec2.ini b/inventory/aws/hosts/ec2.ini index eaab0a410..1f503b8cf 100644 --- a/inventory/aws/hosts/ec2.ini +++ b/inventory/aws/hosts/ec2.ini @@ -24,24 +24,61 @@ regions_exclude = us-gov-west-1,cn-north-1 # This is the normal destination variable to use. If you are running Ansible # from outside EC2, then 'public_dns_name' makes the most sense. If you are # running Ansible from within EC2, then perhaps you want to use the internal -# address, and should set this to 'private_dns_name'. +# address, and should set this to 'private_dns_name'. The key of an EC2 tag +# may optionally be used; however the boto instance variables hold precedence +# in the event of a collision. destination_variable = public_dns_name # For server inside a VPC, using DNS names may not make sense. When an instance # has 'subnet_id' set, this variable is used. If the subnet is public, setting # this to 'ip_address' will return the public IP address. For instances in a # private subnet, this should be set to 'private_ip_address', and Ansible must -# be run from with EC2. +# be run from within EC2. The key of an EC2 tag may optionally be used; however +# the boto instance variables hold precedence in the event of a collision. +# WARNING: - instances that are in the private vpc, _without_ public ip address +# will not be listed in the inventory until You set: +# vpc_destination_variable = 'private_ip_address' vpc_destination_variable = ip_address # To tag instances on EC2 with the resource records that point to them from # Route53, uncomment and set 'route53' to True. route53 = False +# To exclude RDS instances from the inventory, uncomment and set to False. +#rds = False + +# To exclude ElastiCache instances from the inventory, uncomment and set to False. +#elasticache = False + # Additionally, you can specify the list of zones to exclude looking up in # 'route53_excluded_zones' as a comma-separated list. # route53_excluded_zones = samplezone1.com, samplezone2.com +# By default, only EC2 instances in the 'running' state are returned. Set +# 'all_instances' to True to return all instances regardless of state. +all_instances = False + +# By default, only EC2 instances in the 'running' state are returned. Specify +# EC2 instance states to return as a comma-separated list. This +# option is overriden when 'all_instances' is True. +# instance_states = pending, running, shutting-down, terminated, stopping, stopped + +# By default, only RDS instances in the 'available' state are returned. Set +# 'all_rds_instances' to True return all RDS instances regardless of state. +all_rds_instances = False + +# By default, only ElastiCache clusters and nodes in the 'available' state +# are returned. Set 'all_elasticache_clusters' and/or 'all_elastic_nodes' +# to True return all ElastiCache clusters and nodes, regardless of state. +# +# Note that all_elasticache_nodes only applies to listed clusters. That means +# if you set all_elastic_clusters to false, no node will be return from +# unavailable clusters, regardless of the state and to what you set for +# all_elasticache_nodes. +all_elasticache_replication_groups = False +all_elasticache_clusters = False +all_elasticache_nodes = False + # API calls to EC2 are slow. For this reason, we cache the results of an API # call. Set this to the path you want cache files to be written to. Two files # will be written to this directory: @@ -60,3 +97,59 @@ cache_max_age = 300 # destination_variable and vpc_destination_variable. # destination_format = {0}.{1}.rhcloud.com # destination_format_tags = Name,environment + +# Organize groups into a nested/hierarchy instead of a flat namespace. +nested_groups = False + +# Replace - tags when creating groups to avoid issues with ansible +replace_dash_in_groups = False + +# The EC2 inventory output can become very large. To manage its size, +# configure which groups should be created. +group_by_instance_id = True +group_by_region = True +group_by_availability_zone = True +group_by_ami_id = True +group_by_instance_type = True +group_by_key_pair = True +group_by_vpc_id = True +group_by_security_group = True +group_by_tag_keys = True +group_by_tag_none = True +group_by_route53_names = True +group_by_rds_engine = True +group_by_rds_parameter_group = True +group_by_elasticache_engine = True +group_by_elasticache_cluster = True +group_by_elasticache_parameter_group = True +group_by_elasticache_replication_group = True + +# If you only want to include hosts that match a certain regular expression +# pattern_include = staging-* + +# If you want to exclude any hosts that match a certain regular expression +# pattern_exclude = staging-* + +# Instance filters can be used to control which instances are retrieved for +# inventory. For the full list of possible filters, please read the EC2 API +# docs: http://docs.aws.amazon.com/AWSEC2/latest/APIReference/ApiReference-query-DescribeInstances.html#query-DescribeInstances-filters +# Filters are key/value pairs separated by '=', to list multiple filters use +# a list separated by commas. See examples below. + +# Retrieve only instances with (key=value) env=staging tag +# instance_filters = tag:env=staging + +# Retrieve only instances with role=webservers OR role=dbservers tag +# instance_filters = tag:role=webservers,tag:role=dbservers + +# Retrieve only t1.micro instances OR instances with tag env=staging +# instance_filters = instance-type=t1.micro,tag:env=staging + +# You can use wildcards in filter values also. Below will list instances which +# tag Name value matches webservers1* +# (ex. webservers15, webservers1a, webservers123 etc) +# instance_filters = tag:Name=webservers1* + +# A boto configuration profile may be used to separate out credentials +# see http://boto.readthedocs.org/en/latest/boto_config_tut.html +# boto_profile = some-boto-profile-name diff --git a/inventory/aws/hosts/ec2.py b/inventory/aws/hosts/ec2.py index f231ff4c2..8b878cafd 100755 --- a/inventory/aws/hosts/ec2.py +++ b/inventory/aws/hosts/ec2.py @@ -22,6 +22,12 @@ you need to define: export EC2_URL=http://hostname_of_your_cc:port/services/Eucalyptus +If you're using boto profiles (requires boto>=2.24.0) you can choose a profile +using the --boto-profile command line argument (e.g. ec2.py --boto-profile prod) or using +the AWS_PROFILE variable: + + AWS_PROFILE=prod ansible-playbook -i ec2.py myplaybook.yml + For more details, see: http://docs.pythonboto.org/en/latest/boto_config_tut.html When run against a specific host, this script returns the following variables: @@ -121,8 +127,11 @@ from time import time import boto from boto import ec2 from boto import rds +from boto import elasticache from boto import route53 -import ConfigParser +import six + +from six.moves import configparser from collections import defaultdict try: @@ -145,9 +154,18 @@ class Ec2Inventory(object): # Index of hostname (address) to instance ID self.index = {} + # Boto profile to use (if any) + self.boto_profile = None + # Read settings and parse CLI arguments - self.read_settings() self.parse_cli_args() + self.read_settings() + + # Make sure that profile_name is not passed at all if not set + # as pre 2.24 boto will fall over otherwise + if self.boto_profile: + if not hasattr(boto.ec2.EC2Connection, 'profile_name'): + self.fail_with_error("boto version must be >= 2.24 to use profile") # Cache if self.args.refresh_cache: @@ -166,7 +184,7 @@ class Ec2Inventory(object): else: data_to_print = self.json_format_dict(self.inventory, True) - print data_to_print + print(data_to_print) def is_cache_valid(self): @@ -184,10 +202,12 @@ class Ec2Inventory(object): def read_settings(self): ''' Reads the settings from the ec2.ini file ''' - - config = ConfigParser.SafeConfigParser() + if six.PY3: + config = configparser.ConfigParser() + else: + config = configparser.SafeConfigParser() ec2_default_ini_path = os.path.join(os.path.dirname(os.path.realpath(__file__)), 'ec2.ini') - ec2_ini_path = os.environ.get('EC2_INI_PATH', ec2_default_ini_path) + ec2_ini_path = os.path.expanduser(os.path.expandvars(os.environ.get('EC2_INI_PATH', ec2_default_ini_path))) config.read(ec2_ini_path) # is eucalyptus? @@ -236,18 +256,72 @@ class Ec2Inventory(object): if config.has_option('ec2', 'rds'): self.rds_enabled = config.getboolean('ec2', 'rds') - # Return all EC2 and RDS instances (if RDS is enabled) + # Include ElastiCache instances? + self.elasticache_enabled = True + if config.has_option('ec2', 'elasticache'): + self.elasticache_enabled = config.getboolean('ec2', 'elasticache') + + # Return all EC2 instances? if config.has_option('ec2', 'all_instances'): self.all_instances = config.getboolean('ec2', 'all_instances') else: self.all_instances = False + + # Instance states to be gathered in inventory. Default is 'running'. + # Setting 'all_instances' to 'yes' overrides this option. + ec2_valid_instance_states = [ + 'pending', + 'running', + 'shutting-down', + 'terminated', + 'stopping', + 'stopped' + ] + self.ec2_instance_states = [] + if self.all_instances: + self.ec2_instance_states = ec2_valid_instance_states + elif config.has_option('ec2', 'instance_states'): + for instance_state in config.get('ec2', 'instance_states').split(','): + instance_state = instance_state.strip() + if instance_state not in ec2_valid_instance_states: + continue + self.ec2_instance_states.append(instance_state) + else: + self.ec2_instance_states = ['running'] + + # Return all RDS instances? (if RDS is enabled) if config.has_option('ec2', 'all_rds_instances') and self.rds_enabled: self.all_rds_instances = config.getboolean('ec2', 'all_rds_instances') else: self.all_rds_instances = False + # Return all ElastiCache replication groups? (if ElastiCache is enabled) + if config.has_option('ec2', 'all_elasticache_replication_groups') and self.elasticache_enabled: + self.all_elasticache_replication_groups = config.getboolean('ec2', 'all_elasticache_replication_groups') + else: + self.all_elasticache_replication_groups = False + + # Return all ElastiCache clusters? (if ElastiCache is enabled) + if config.has_option('ec2', 'all_elasticache_clusters') and self.elasticache_enabled: + self.all_elasticache_clusters = config.getboolean('ec2', 'all_elasticache_clusters') + else: + self.all_elasticache_clusters = False + + # Return all ElastiCache nodes? (if ElastiCache is enabled) + if config.has_option('ec2', 'all_elasticache_nodes') and self.elasticache_enabled: + self.all_elasticache_nodes = config.getboolean('ec2', 'all_elasticache_nodes') + else: + self.all_elasticache_nodes = False + + # boto configuration profile (prefer CLI argument) + self.boto_profile = self.args.boto_profile + if config.has_option('ec2', 'boto_profile') and not self.boto_profile: + self.boto_profile = config.get('ec2', 'boto_profile') + # Cache related cache_dir = os.path.expanduser(config.get('ec2', 'cache_path')) + if self.boto_profile: + cache_dir = os.path.join(cache_dir, 'profile_' + self.boto_profile) if not os.path.exists(cache_dir): os.makedirs(cache_dir) @@ -261,6 +335,12 @@ class Ec2Inventory(object): else: self.nested_groups = False + # Replace dash or not in group names + if config.has_option('ec2', 'replace_dash_in_groups'): + self.replace_dash_in_groups = config.getboolean('ec2', 'replace_dash_in_groups') + else: + self.replace_dash_in_groups = True + # Configure which groups should be created. group_by_options = [ 'group_by_instance_id', @@ -276,6 +356,10 @@ class Ec2Inventory(object): 'group_by_route53_names', 'group_by_rds_engine', 'group_by_rds_parameter_group', + 'group_by_elasticache_engine', + 'group_by_elasticache_cluster', + 'group_by_elasticache_parameter_group', + 'group_by_elasticache_replication_group', ] for option in group_by_options: if config.has_option('ec2', option): @@ -290,7 +374,7 @@ class Ec2Inventory(object): self.pattern_include = re.compile(pattern_include) else: self.pattern_include = None - except ConfigParser.NoOptionError, e: + except configparser.NoOptionError: self.pattern_include = None # Do we need to exclude hosts that match a pattern? @@ -300,7 +384,7 @@ class Ec2Inventory(object): self.pattern_exclude = re.compile(pattern_exclude) else: self.pattern_exclude = None - except ConfigParser.NoOptionError, e: + except configparser.NoOptionError: self.pattern_exclude = None # Instance filters (see boto and EC2 API docs). Ignore invalid filters. @@ -325,6 +409,8 @@ class Ec2Inventory(object): help='Get all the variables about a specific instance') parser.add_argument('--refresh-cache', action='store_true', default=False, help='Force refresh of cache by making API requests to EC2 (default: False - use cache files)') + parser.add_argument('--boto-profile', action='store', + help='Use boto profile for connections to EC2') self.args = parser.parse_args() @@ -338,30 +424,52 @@ class Ec2Inventory(object): self.get_instances_by_region(region) if self.rds_enabled: self.get_rds_instances_by_region(region) + if self.elasticache_enabled: + self.get_elasticache_clusters_by_region(region) + self.get_elasticache_replication_groups_by_region(region) self.write_to_cache(self.inventory, self.cache_path_cache) self.write_to_cache(self.index, self.cache_path_index) + def connect(self, region): + ''' create connection to api server''' + if self.eucalyptus: + conn = boto.connect_euca(host=self.eucalyptus_host) + conn.APIVersion = '2010-08-31' + else: + conn = self.connect_to_aws(ec2, region) + return conn + + def boto_fix_security_token_in_profile(self, connect_args): + ''' monkey patch for boto issue boto/boto#2100 ''' + profile = 'profile ' + self.boto_profile + if boto.config.has_option(profile, 'aws_security_token'): + connect_args['security_token'] = boto.config.get(profile, 'aws_security_token') + return connect_args + + def connect_to_aws(self, module, region): + connect_args = {} + + # only pass the profile name if it's set (as it is not supported by older boto versions) + if self.boto_profile: + connect_args['profile_name'] = self.boto_profile + self.boto_fix_security_token_in_profile(connect_args) + + conn = module.connect_to_region(region, **connect_args) + # connect_to_region will fail "silently" by returning None if the region name is wrong or not supported + if conn is None: + self.fail_with_error("region name: %s likely not supported, or AWS is down. connection to region failed." % region) + return conn def get_instances_by_region(self, region): ''' Makes an AWS EC2 API call to the list of instances in a particular region ''' try: - if self.eucalyptus: - conn = boto.connect_euca(host=self.eucalyptus_host) - conn.APIVersion = '2010-08-31' - else: - conn = ec2.connect_to_region(region) - - # connect_to_region will fail "silently" by returning None if the region name is wrong or not supported - if conn is None: - print("region name: %s likely not supported, or AWS is down. connection to region failed." % region) - sys.exit(1) - + conn = self.connect(region) reservations = [] if self.ec2_instance_filters: - for filter_key, filter_values in self.ec2_instance_filters.iteritems(): + for filter_key, filter_values in self.ec2_instance_filters.items(): reservations.extend(conn.get_all_instances(filters = { filter_key : filter_values })) else: reservations = conn.get_all_instances() @@ -370,40 +478,130 @@ class Ec2Inventory(object): for instance in reservation.instances: self.add_instance(instance, region) - except boto.exception.BotoServerError, e: - if not self.eucalyptus: - print "Looks like AWS is down again:" - print e - sys.exit(1) + except boto.exception.BotoServerError as e: + if e.error_code == 'AuthFailure': + error = self.get_auth_error_message() + else: + backend = 'Eucalyptus' if self.eucalyptus else 'AWS' + error = "Error connecting to %s backend.\n%s" % (backend, e.message) + self.fail_with_error(error, 'getting EC2 instances') def get_rds_instances_by_region(self, region): ''' Makes an AWS API call to the list of RDS instances in a particular region ''' try: - conn = rds.connect_to_region(region) + conn = self.connect_to_aws(rds, region) if conn: instances = conn.get_all_dbinstances() for instance in instances: self.add_rds_instance(instance, region) - except boto.exception.BotoServerError, e: + except boto.exception.BotoServerError as e: + error = e.reason + + if e.error_code == 'AuthFailure': + error = self.get_auth_error_message() if not e.reason == "Forbidden": - print "Looks like AWS RDS is down: " - print e - sys.exit(1) + error = "Looks like AWS RDS is down:\n%s" % e.message + self.fail_with_error(error, 'getting RDS instances') - def get_instance(self, region, instance_id): - ''' Gets details about a specific instance ''' - if self.eucalyptus: - conn = boto.connect_euca(self.eucalyptus_host) - conn.APIVersion = '2010-08-31' + def get_elasticache_clusters_by_region(self, region): + ''' Makes an AWS API call to the list of ElastiCache clusters (with + nodes' info) in a particular region.''' + + # ElastiCache boto module doesn't provide a get_all_intances method, + # that's why we need to call describe directly (it would be called by + # the shorthand method anyway...) + try: + conn = elasticache.connect_to_region(region) + if conn: + # show_cache_node_info = True + # because we also want nodes' information + response = conn.describe_cache_clusters(None, None, None, True) + + except boto.exception.BotoServerError as e: + error = e.reason + + if e.error_code == 'AuthFailure': + error = self.get_auth_error_message() + if not e.reason == "Forbidden": + error = "Looks like AWS ElastiCache is down:\n%s" % e.message + self.fail_with_error(error, 'getting ElastiCache clusters') + + try: + # Boto also doesn't provide wrapper classes to CacheClusters or + # CacheNodes. Because of that wo can't make use of the get_list + # method in the AWSQueryConnection. Let's do the work manually + clusters = response['DescribeCacheClustersResponse']['DescribeCacheClustersResult']['CacheClusters'] + + except KeyError as e: + error = "ElastiCache query to AWS failed (unexpected format)." + self.fail_with_error(error, 'getting ElastiCache clusters') + + for cluster in clusters: + self.add_elasticache_cluster(cluster, region) + + def get_elasticache_replication_groups_by_region(self, region): + ''' Makes an AWS API call to the list of ElastiCache replication groups + in a particular region.''' + + # ElastiCache boto module doesn't provide a get_all_intances method, + # that's why we need to call describe directly (it would be called by + # the shorthand method anyway...) + try: + conn = elasticache.connect_to_region(region) + if conn: + response = conn.describe_replication_groups() + + except boto.exception.BotoServerError as e: + error = e.reason + + if e.error_code == 'AuthFailure': + error = self.get_auth_error_message() + if not e.reason == "Forbidden": + error = "Looks like AWS ElastiCache [Replication Groups] is down:\n%s" % e.message + self.fail_with_error(error, 'getting ElastiCache clusters') + + try: + # Boto also doesn't provide wrapper classes to ReplicationGroups + # Because of that wo can't make use of the get_list method in the + # AWSQueryConnection. Let's do the work manually + replication_groups = response['DescribeReplicationGroupsResponse']['DescribeReplicationGroupsResult']['ReplicationGroups'] + + except KeyError as e: + error = "ElastiCache [Replication Groups] query to AWS failed (unexpected format)." + self.fail_with_error(error, 'getting ElastiCache clusters') + + for replication_group in replication_groups: + self.add_elasticache_replication_group(replication_group, region) + + def get_auth_error_message(self): + ''' create an informative error message if there is an issue authenticating''' + errors = ["Authentication error retrieving ec2 inventory."] + if None in [os.environ.get('AWS_ACCESS_KEY_ID'), os.environ.get('AWS_SECRET_ACCESS_KEY')]: + errors.append(' - No AWS_ACCESS_KEY_ID or AWS_SECRET_ACCESS_KEY environment vars found') else: - conn = ec2.connect_to_region(region) + errors.append(' - AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY environment vars found but may not be correct') - # connect_to_region will fail "silently" by returning None if the region name is wrong or not supported - if conn is None: - print("region name: %s likely not supported, or AWS is down. connection to region failed." % region) - sys.exit(1) + boto_paths = ['/etc/boto.cfg', '~/.boto', '~/.aws/credentials'] + boto_config_found = list(p for p in boto_paths if os.path.isfile(os.path.expanduser(p))) + if len(boto_config_found) > 0: + errors.append(" - Boto configs found at '%s', but the credentials contained may not be correct" % ', '.join(boto_config_found)) + else: + errors.append(" - No Boto config found at any expected location '%s'" % ', '.join(boto_paths)) + + return '\n'.join(errors) + + def fail_with_error(self, err_msg, err_operation=None): + '''log an error to std err for ansible-playbook to consume and exit''' + if err_operation: + err_msg = 'ERROR: "{err_msg}", while: {err_operation}'.format( + err_msg=err_msg, err_operation=err_operation) + sys.stderr.write(err_msg) + sys.exit(1) + + def get_instance(self, region, instance_id): + conn = self.connect(region) reservations = conn.get_all_instances([instance_id]) for reservation in reservations: @@ -414,8 +612,8 @@ class Ec2Inventory(object): ''' Adds an instance to the inventory and index, as long as it is addressable ''' - # Only want running instances unless all_instances is True - if not self.all_instances and instance.state != 'running': + # Only return instances with desired instance states + if instance.state not in self.ec2_instance_states: return # Select the best destination address @@ -502,18 +700,21 @@ class Ec2Inventory(object): if self.nested_groups: self.push_group(self.inventory, 'security_groups', key) except AttributeError: - print 'Package boto seems a bit older.' - print 'Please upgrade boto >= 2.3.0.' - sys.exit(1) + self.fail_with_error('\n'.join(['Package boto seems a bit older.', + 'Please upgrade boto >= 2.3.0.'])) # Inventory: Group by tag keys if self.group_by_tag_keys: - for k, v in instance.tags.iteritems(): - key = self.to_safe("tag_" + k + "=" + v) + for k, v in instance.tags.items(): + if v: + key = self.to_safe("tag_" + k + "=" + v) + else: + key = self.to_safe("tag_" + k) self.push(self.inventory, key, dest) if self.nested_groups: self.push_group(self.inventory, 'tags', self.to_safe("tag_" + k)) - self.push_group(self.inventory, self.to_safe("tag_" + k), key) + if v: + self.push_group(self.inventory, self.to_safe("tag_" + k), key) # Inventory: Group by Route53 domain names if enabled if self.route53_enabled and self.group_by_route53_names: @@ -597,9 +798,9 @@ class Ec2Inventory(object): self.push_group(self.inventory, 'security_groups', key) except AttributeError: - print 'Package boto seems a bit older.' - print 'Please upgrade boto >= 2.3.0.' - sys.exit(1) + self.fail_with_error('\n'.join(['Package boto seems a bit older.', + 'Please upgrade boto >= 2.3.0.'])) + # Inventory: Group by engine if self.group_by_rds_engine: @@ -618,6 +819,243 @@ class Ec2Inventory(object): self.inventory["_meta"]["hostvars"][dest] = self.get_host_info_dict_from_instance(instance) + def add_elasticache_cluster(self, cluster, region): + ''' Adds an ElastiCache cluster to the inventory and index, as long as + it's nodes are addressable ''' + + # Only want available clusters unless all_elasticache_clusters is True + if not self.all_elasticache_clusters and cluster['CacheClusterStatus'] != 'available': + return + + # Select the best destination address + if 'ConfigurationEndpoint' in cluster and cluster['ConfigurationEndpoint']: + # Memcached cluster + dest = cluster['ConfigurationEndpoint']['Address'] + is_redis = False + else: + # Redis sigle node cluster + # Because all Redis clusters are single nodes, we'll merge the + # info from the cluster with info about the node + dest = cluster['CacheNodes'][0]['Endpoint']['Address'] + is_redis = True + + if not dest: + # Skip clusters we cannot address (e.g. private VPC subnet) + return + + # Add to index + self.index[dest] = [region, cluster['CacheClusterId']] + + # Inventory: Group by instance ID (always a group of 1) + if self.group_by_instance_id: + self.inventory[cluster['CacheClusterId']] = [dest] + if self.nested_groups: + self.push_group(self.inventory, 'instances', cluster['CacheClusterId']) + + # Inventory: Group by region + if self.group_by_region and not is_redis: + self.push(self.inventory, region, dest) + if self.nested_groups: + self.push_group(self.inventory, 'regions', region) + + # Inventory: Group by availability zone + if self.group_by_availability_zone and not is_redis: + self.push(self.inventory, cluster['PreferredAvailabilityZone'], dest) + if self.nested_groups: + if self.group_by_region: + self.push_group(self.inventory, region, cluster['PreferredAvailabilityZone']) + self.push_group(self.inventory, 'zones', cluster['PreferredAvailabilityZone']) + + # Inventory: Group by node type + if self.group_by_instance_type and not is_redis: + type_name = self.to_safe('type_' + cluster['CacheNodeType']) + self.push(self.inventory, type_name, dest) + if self.nested_groups: + self.push_group(self.inventory, 'types', type_name) + + # Inventory: Group by VPC (information not available in the current + # AWS API version for ElastiCache) + + # Inventory: Group by security group + if self.group_by_security_group and not is_redis: + + # Check for the existence of the 'SecurityGroups' key and also if + # this key has some value. When the cluster is not placed in a SG + # the query can return None here and cause an error. + if 'SecurityGroups' in cluster and cluster['SecurityGroups'] is not None: + for security_group in cluster['SecurityGroups']: + key = self.to_safe("security_group_" + security_group['SecurityGroupId']) + self.push(self.inventory, key, dest) + if self.nested_groups: + self.push_group(self.inventory, 'security_groups', key) + + # Inventory: Group by engine + if self.group_by_elasticache_engine and not is_redis: + self.push(self.inventory, self.to_safe("elasticache_" + cluster['Engine']), dest) + if self.nested_groups: + self.push_group(self.inventory, 'elasticache_engines', self.to_safe(cluster['Engine'])) + + # Inventory: Group by parameter group + if self.group_by_elasticache_parameter_group: + self.push(self.inventory, self.to_safe("elasticache_parameter_group_" + cluster['CacheParameterGroup']['CacheParameterGroupName']), dest) + if self.nested_groups: + self.push_group(self.inventory, 'elasticache_parameter_groups', self.to_safe(cluster['CacheParameterGroup']['CacheParameterGroupName'])) + + # Inventory: Group by replication group + if self.group_by_elasticache_replication_group and 'ReplicationGroupId' in cluster and cluster['ReplicationGroupId']: + self.push(self.inventory, self.to_safe("elasticache_replication_group_" + cluster['ReplicationGroupId']), dest) + if self.nested_groups: + self.push_group(self.inventory, 'elasticache_replication_groups', self.to_safe(cluster['ReplicationGroupId'])) + + # Global Tag: all ElastiCache clusters + self.push(self.inventory, 'elasticache_clusters', cluster['CacheClusterId']) + + host_info = self.get_host_info_dict_from_describe_dict(cluster) + + self.inventory["_meta"]["hostvars"][dest] = host_info + + # Add the nodes + for node in cluster['CacheNodes']: + self.add_elasticache_node(node, cluster, region) + + def add_elasticache_node(self, node, cluster, region): + ''' Adds an ElastiCache node to the inventory and index, as long as + it is addressable ''' + + # Only want available nodes unless all_elasticache_nodes is True + if not self.all_elasticache_nodes and node['CacheNodeStatus'] != 'available': + return + + # Select the best destination address + dest = node['Endpoint']['Address'] + + if not dest: + # Skip nodes we cannot address (e.g. private VPC subnet) + return + + node_id = self.to_safe(cluster['CacheClusterId'] + '_' + node['CacheNodeId']) + + # Add to index + self.index[dest] = [region, node_id] + + # Inventory: Group by node ID (always a group of 1) + if self.group_by_instance_id: + self.inventory[node_id] = [dest] + if self.nested_groups: + self.push_group(self.inventory, 'instances', node_id) + + # Inventory: Group by region + if self.group_by_region: + self.push(self.inventory, region, dest) + if self.nested_groups: + self.push_group(self.inventory, 'regions', region) + + # Inventory: Group by availability zone + if self.group_by_availability_zone: + self.push(self.inventory, cluster['PreferredAvailabilityZone'], dest) + if self.nested_groups: + if self.group_by_region: + self.push_group(self.inventory, region, cluster['PreferredAvailabilityZone']) + self.push_group(self.inventory, 'zones', cluster['PreferredAvailabilityZone']) + + # Inventory: Group by node type + if self.group_by_instance_type: + type_name = self.to_safe('type_' + cluster['CacheNodeType']) + self.push(self.inventory, type_name, dest) + if self.nested_groups: + self.push_group(self.inventory, 'types', type_name) + + # Inventory: Group by VPC (information not available in the current + # AWS API version for ElastiCache) + + # Inventory: Group by security group + if self.group_by_security_group: + + # Check for the existence of the 'SecurityGroups' key and also if + # this key has some value. When the cluster is not placed in a SG + # the query can return None here and cause an error. + if 'SecurityGroups' in cluster and cluster['SecurityGroups'] is not None: + for security_group in cluster['SecurityGroups']: + key = self.to_safe("security_group_" + security_group['SecurityGroupId']) + self.push(self.inventory, key, dest) + if self.nested_groups: + self.push_group(self.inventory, 'security_groups', key) + + # Inventory: Group by engine + if self.group_by_elasticache_engine: + self.push(self.inventory, self.to_safe("elasticache_" + cluster['Engine']), dest) + if self.nested_groups: + self.push_group(self.inventory, 'elasticache_engines', self.to_safe("elasticache_" + cluster['Engine'])) + + # Inventory: Group by parameter group (done at cluster level) + + # Inventory: Group by replication group (done at cluster level) + + # Inventory: Group by ElastiCache Cluster + if self.group_by_elasticache_cluster: + self.push(self.inventory, self.to_safe("elasticache_cluster_" + cluster['CacheClusterId']), dest) + + # Global Tag: all ElastiCache nodes + self.push(self.inventory, 'elasticache_nodes', dest) + + host_info = self.get_host_info_dict_from_describe_dict(node) + + if dest in self.inventory["_meta"]["hostvars"]: + self.inventory["_meta"]["hostvars"][dest].update(host_info) + else: + self.inventory["_meta"]["hostvars"][dest] = host_info + + def add_elasticache_replication_group(self, replication_group, region): + ''' Adds an ElastiCache replication group to the inventory and index ''' + + # Only want available clusters unless all_elasticache_replication_groups is True + if not self.all_elasticache_replication_groups and replication_group['Status'] != 'available': + return + + # Select the best destination address (PrimaryEndpoint) + dest = replication_group['NodeGroups'][0]['PrimaryEndpoint']['Address'] + + if not dest: + # Skip clusters we cannot address (e.g. private VPC subnet) + return + + # Add to index + self.index[dest] = [region, replication_group['ReplicationGroupId']] + + # Inventory: Group by ID (always a group of 1) + if self.group_by_instance_id: + self.inventory[replication_group['ReplicationGroupId']] = [dest] + if self.nested_groups: + self.push_group(self.inventory, 'instances', replication_group['ReplicationGroupId']) + + # Inventory: Group by region + if self.group_by_region: + self.push(self.inventory, region, dest) + if self.nested_groups: + self.push_group(self.inventory, 'regions', region) + + # Inventory: Group by availability zone (doesn't apply to replication groups) + + # Inventory: Group by node type (doesn't apply to replication groups) + + # Inventory: Group by VPC (information not available in the current + # AWS API version for replication groups + + # Inventory: Group by security group (doesn't apply to replication groups) + # Check this value in cluster level + + # Inventory: Group by engine (replication groups are always Redis) + if self.group_by_elasticache_engine: + self.push(self.inventory, 'elasticache_redis', dest) + if self.nested_groups: + self.push_group(self.inventory, 'elasticache_engines', 'redis') + + # Global Tag: all ElastiCache clusters + self.push(self.inventory, 'elasticache_replication_groups', replication_group['ReplicationGroupId']) + + host_info = self.get_host_info_dict_from_describe_dict(replication_group) + + self.inventory["_meta"]["hostvars"][dest] = host_info def get_route53_records(self): ''' Get and store the map of resource records to domain names that @@ -666,7 +1104,6 @@ class Ec2Inventory(object): return list(name_list) - def get_host_info_dict_from_instance(self, instance): instance_vars = {} for key in vars(instance): @@ -683,7 +1120,7 @@ class Ec2Inventory(object): instance_vars['ec2_previous_state_code'] = instance.previous_state_code elif type(value) in [int, bool]: instance_vars[key] = value - elif type(value) in [str, unicode]: + elif isinstance(value, six.string_types): instance_vars[key] = value.strip() elif type(value) == type(None): instance_vars[key] = '' @@ -692,7 +1129,7 @@ class Ec2Inventory(object): elif key == 'ec2__placement': instance_vars['ec2_placement'] = value.zone elif key == 'ec2_tags': - for k, v in value.iteritems(): + for k, v in value.items(): key = self.to_safe('ec2_tag_' + k) instance_vars[key] = v elif key == 'ec2_groups': @@ -712,6 +1149,91 @@ class Ec2Inventory(object): return instance_vars + def get_host_info_dict_from_describe_dict(self, describe_dict): + ''' Parses the dictionary returned by the API call into a flat list + of parameters. This method should be used only when 'describe' is + used directly because Boto doesn't provide specific classes. ''' + + # I really don't agree with prefixing everything with 'ec2' + # because EC2, RDS and ElastiCache are different services. + # I'm just following the pattern used until now to not break any + # compatibility. + + host_info = {} + for key in describe_dict: + value = describe_dict[key] + key = self.to_safe('ec2_' + self.uncammelize(key)) + + # Handle complex types + + # Target: Memcached Cache Clusters + if key == 'ec2_configuration_endpoint' and value: + host_info['ec2_configuration_endpoint_address'] = value['Address'] + host_info['ec2_configuration_endpoint_port'] = value['Port'] + + # Target: Cache Nodes and Redis Cache Clusters (single node) + if key == 'ec2_endpoint' and value: + host_info['ec2_endpoint_address'] = value['Address'] + host_info['ec2_endpoint_port'] = value['Port'] + + # Target: Redis Replication Groups + if key == 'ec2_node_groups' and value: + host_info['ec2_endpoint_address'] = value[0]['PrimaryEndpoint']['Address'] + host_info['ec2_endpoint_port'] = value[0]['PrimaryEndpoint']['Port'] + replica_count = 0 + for node in value[0]['NodeGroupMembers']: + if node['CurrentRole'] == 'primary': + host_info['ec2_primary_cluster_address'] = node['ReadEndpoint']['Address'] + host_info['ec2_primary_cluster_port'] = node['ReadEndpoint']['Port'] + host_info['ec2_primary_cluster_id'] = node['CacheClusterId'] + elif node['CurrentRole'] == 'replica': + host_info['ec2_replica_cluster_address_'+ str(replica_count)] = node['ReadEndpoint']['Address'] + host_info['ec2_replica_cluster_port_'+ str(replica_count)] = node['ReadEndpoint']['Port'] + host_info['ec2_replica_cluster_id_'+ str(replica_count)] = node['CacheClusterId'] + replica_count += 1 + + # Target: Redis Replication Groups + if key == 'ec2_member_clusters' and value: + host_info['ec2_member_clusters'] = ','.join([str(i) for i in value]) + + # Target: All Cache Clusters + elif key == 'ec2_cache_parameter_group': + host_info["ec2_cache_node_ids_to_reboot"] = ','.join([str(i) for i in value['CacheNodeIdsToReboot']]) + host_info['ec2_cache_parameter_group_name'] = value['CacheParameterGroupName'] + host_info['ec2_cache_parameter_apply_status'] = value['ParameterApplyStatus'] + + # Target: Almost everything + elif key == 'ec2_security_groups': + + # Skip if SecurityGroups is None + # (it is possible to have the key defined but no value in it). + if value is not None: + sg_ids = [] + for sg in value: + sg_ids.append(sg['SecurityGroupId']) + host_info["ec2_security_group_ids"] = ','.join([str(i) for i in sg_ids]) + + # Target: Everything + # Preserve booleans and integers + elif type(value) in [int, bool]: + host_info[key] = value + + # Target: Everything + # Sanitize string values + elif isinstance(value, six.string_types): + host_info[key] = value.strip() + + # Target: Everything + # Replace None by an empty string + elif type(value) == type(None): + host_info[key] = '' + + else: + # Remove non-processed complex types + pass + + return host_info + def get_host_info(self): ''' Get variables about a specific host ''' @@ -775,13 +1297,16 @@ class Ec2Inventory(object): cache.write(json_data) cache.close() + def uncammelize(self, key): + temp = re.sub('(.)([A-Z][a-z]+)', r'\1_\2', key) + return re.sub('([a-z0-9])([A-Z])', r'\1_\2', temp).lower() def to_safe(self, word): - ''' Converts 'bad' characters in a string to underscores so they can be - used as Ansible groups ''' - - return re.sub("[^A-Za-z0-9\-]", "_", word) - + ''' Converts 'bad' characters in a string to underscores so they can be used as Ansible groups ''' + regex = "[^A-Za-z0-9\_" + if not self.replace_dash_in_groups: + regex += "\-" + return re.sub(regex + "]", "_", word) def json_format_dict(self, data, pretty=False): ''' Converts a dict to a JSON object and dumps it as a formatted -- cgit v1.2.1