Allow ec2 inventory to use a boto profile

This allows the EC2 inventory plugin to be used with
the same configuration against different EC2 accounts

Profile can be passed using --profile variable or using
EC2_PROFILE environment variable e.g.

```
EC2_PROFILE=prod ansible-playbook -i ec2.py playbook.yml
```

Added documentation on profiles to EC2 dynamic inventory doc

Only tries to use profiles if --profile argument is given
or EC2_PROFILE is set to maintain compatibility will boto < 2.24.

Works around a minor bug in boto where if you try and use
a security token with a profile it fails (boto/boto#2100)

contrib/inventory/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 --profile command line argument (e.g. ec2.py --profile prod) or using
+the EC2_PROFILE variable:
+
+    EC2_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:
@@ -148,9 +154,20 @@ class Ec2Inventory(object):
         # Index of hostname (address) to instance ID
         self.index = {}
 
-        # Read settings and parse CLI arguments
-        self.read_settings()
+        # Parse CLI arguments and read settings
         self.parse_cli_args()
+        self.read_settings()
+
+        # boto profile to use (if any)
+        # Make sure that profile_name is not passed at all if not set
+        # as pre 2.24 boto will fall over otherwise
+        if self.args.profile:
+            if not hasattr(boto.ec2.EC2Connection, 'profile_name'):
+                sys.stderr.write("boto version must be >= 2.24 to use profile\n")
+                sys.exit(1)
+            self.profile = dict(profile_name=self.args.profile)
+        else:
+            self.profile = dict()
 
         # Cache
         if self.args.refresh_cache:
@@ -292,6 +309,8 @@ class Ec2Inventory(object):
 
         # Cache related
         cache_dir = os.path.expanduser(config.get('ec2', 'cache_path'))
+        if self.args.profile:
+            cache_dir = os.path.join(cache_dir, 'profile_' + self.args.profile)
         if not os.path.exists(cache_dir):
             os.makedirs(cache_dir)
 
@@ -373,6 +392,8 @@ 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('--profile', action='store', default=os.environ.get('EC2_PROFILE'),
+                           help='Use boto profile for connections to EC2')
         self.args = parser.parse_args()
 
 
@@ -405,6 +426,21 @@ class Ec2Inventory(object):
             self.fail_with_error("region name: %s likely not supported, or AWS is down.  connection to region failed." % region)
         return conn
 
+    def boto_fix_security_token_in_profile(self, conn):
+        ''' monkey patch for boto issue boto/boto#2100 '''
+        profile = 'profile ' + self.profile.get('profile_name')
+        if boto.config.has_option(profile, 'aws_security_token'):
+            conn.provider.set_security_token(boto.config.get(profile, 'aws_security_token'))
+        return conn
+
+
+    def connect_to_aws(self, module, region):
+        conn = module.connect_to_region(region, **self.profile)
+        if 'profile_name' in self.profile:
+            conn = self.boto_fix_security_token_in_profile(conn)
+        return conn
+
+
     def get_instances_by_region(self, region):
         ''' Makes an AWS EC2 API call to the list of instances in a particular
         region '''
@@ -416,8 +452,14 @@ class Ec2Inventory(object):
                 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()
+                conn = self.connect_to_aws(ec2, 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)
+
+            reservations = conn.get_all_instances()
             for reservation in reservations:
                 for instance in reservation.instances:
                     self.add_instance(instance, region)
@@ -430,12 +472,13 @@ class Ec2Inventory(object):
                 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:

docsite/rst/intro_dynamic_inventory.rst

@@ -101,6 +101,20 @@ You can test the script by itself to make sure your config is correct::
 
 After a few moments, you should see your entire EC2 inventory across all regions in JSON.
 
+If you use boto profiles to manage multiple AWS accounts, you can pass ``--profile PROFILE`` name to the ``ec2.py`` script. An example profile might be::
+
+    [profile dev]
+    aws_access_key_id = <dev access key>
+    aws_secret_access_key = <dev secret key>
+
+    [profile prod]
+    aws_access_key_id = <prod access key>
+    aws_secret_access_key = <prod secret key>
+
+You can then run ``ec2.py --profile prod`` to get the inventory for the prod account, or run playbooks with: ``ansible-playbook -i 'ec2.py --profile prod' myplaybook.yml``.
+
+Alternatively, use the ``EC2_PROFILE`` variable - e.g. ``EC2_PROFILE=prod ansible-playbook -i ec2.py myplaybook.yml``
+
 Since each region requires its own API call, if you are only using a small set of regions, feel free to edit ``ec2.ini`` and list only the regions you are interested in. There are other config options in ``ec2.ini`` including cache control, and destination variables.
 
 At their heart, inventory files are simply a mapping from some name to a destination address. The default ``ec2.ini`` settings are configured for running Ansible from outside EC2 (from your laptop for example) -- and this is not the most efficient way to manage EC2.