dynamic_inventory_cm_py3

#!/usr/bin/env python3
# -*- coding: utf-8 -*-

'''
Created by Gabor Roczei <roczei@cloudera.com>

=============================================================================
Ansible Dynamic Inventory Script for Cloudera Manager
=============================================================================

The available Cloudera Manager clusters (Ansible groups, for example: Cluster_1, Balaton)
can be listed with the following command:

dynamic_inventory_cm --list

Example Ansible ad-hoc commands:

ansible all -u $USER --become-user $USER -i dynamic_inventory_cm --list-hosts
ansible Balaton --ask-pass -u $USER --become-user $USER -i dynamic_inventory_cm -o -m command -a "time id -Gn $USER"
ansible Cluster_1 --ask-pass -u $USER --become-user $USER -i dynamic_inventory_cm -o -m command -a "date"
ansible --ask-pass -u $USER --become-user $USER -i dynamic_inventory_cm -o -m ping all
'''

import os
import time
import sys
import argparse
import json
import urllib.request, urllib.error, urllib.parse
import ssl
import base64
import getpass
import http.cookiejar
from urllib.parse import urlparse

# Cache time (seconds)
default_cache_time_sec = 30 * 24 * 60 * 60

# Disabling SSL CA check
default_disable_ca_check = False

# Cloudera Manager API timeout (seconds)
default_timeout_sec = 60

# Debug mode
default_debug = False

def save_cookie(cookiejar, filename):
    lwp_cookiejar = http.cookiejar.LWPCookieJar()
    for cookie in cookiejar:
        parameters = dict(list(vars(cookie).items()))
        parameters["rest"] = parameters["_rest"]
        del parameters["_rest"]
        cookie = http.cookiejar.Cookie(**parameters)
        lwp_cookiejar.set_cookie(cookie)
    lwp_cookiejar.save(filename, ignore_discard = True)
    os.chmod(filename, 0o600)

def load_cookie(filename):
    lwp_cookiejar = http.cookiejar.LWPCookieJar()
    lwp_cookiejar.load(filename, ignore_discard = True)
    return lwp_cookiejar

class ClouderaManagerInventory:

    def __init__(self):
        self.inventory = {"_meta" : { "hostvars": {} }}
        self.cookie_path = os.environ["HOME"] + os.sep + ".cm" + os.sep + ".cm_cookie"
        self.json_path = os.environ["HOME"] + os.sep + ".cm" + os.sep + ".cm_inventory.json"
        self.api_version = ""

        cache_time_sec = os.getenv("CM_CACHE_TIME_SEC")
        disable_ca_check = os.getenv("CM_DISABLE_CA_CHECK")
        timeout_sec = os.getenv("CM_TIMEOUT_SEC")
        debug = os.getenv("CM_DEBUG")

        if not os.path.exists(os.environ["HOME"] + os.sep + ".cm"):
           os.makedirs(os.environ["HOME"] + os.sep + ".cm")
           os.chmod(os.environ["HOME"] + os.sep + ".cm", 0o700)

        if cache_time_sec == None:
            self.cache_time_sec = default_cache_time_sec
        else:
            try:
               self.cache_time_sec = int(cache_time_sec)
            except Exception as e:
               if "invalid literal for int" in str(e):
                  sys.stderr.write("The value is incorrect in CM_CACHE_TIME_SEC: " + cache_time_sec  + ". It is not a valid integer number.\n")
                  sys.exit(-1)
               else:
                  sys.stderr.write(str(e))
                  sys.exit(-1)

        if timeout_sec == None:
            self.timeout_sec = default_timeout_sec
        else:
            try:
               self.timeout_sec = int(timeout_sec)
            except Exception as e:
               if "invalid literal for int" in str(e):
                  sys.stderr.write("The value is incorrect in CM_TIMEOUT_SEC: " + timeout_sec  + ". It is not a valid integer number.\n")
                  sys.exit(-1)
               else:
                  sys.stderr.write(str(e))
                  sys.exit(-1)
        
        if disable_ca_check == None:
            self.disable_ca_check = default_disable_ca_check
        elif disable_ca_check.lower() == "true":
            self.disable_ca_check = True
        elif disable_ca_check.lower() == "false":
            self.disable_ca_check = False
        else:
            sys.stderr.write("Unknown value in CM_DISABLE_CA_CHECK: " + disable_ca_check  + ". Valid values can only be True or False\n")
            sys.exit(-1)

        if debug == None:
            self.debug = default_debug
        elif debug.lower() == "true":
            self.debug = True
        elif debug.lower() == "false":
            self.debug = False
        else:
            sys.stderr.write("Unknown value in CM_DEBUG: " + debug + ". Valid values can only be True or False\n")
            sys.exit(-1)

        self.read_args()

    def cm_connection(self, url, username, password, cookie):
        try:
             if self.disable_ca_check:
                context = ssl._create_unverified_context()
             else:
                context = None
             
             cookie_handler = urllib.request.HTTPCookieProcessor(cookie)
            
             password_manager = urllib.request.HTTPPasswordMgrWithDefaultRealm()
             password_manager.add_password(realm = "Cloudera Manager",
                          uri = url + "/api/v1",
                          user = username,
                          passwd = password)

             auth_handler = urllib.request.HTTPBasicAuthHandler(password_manager)

             if self.debug:
                debuglevel = 1
             else:
                debuglevel = 0

             http_handler = urllib.request.HTTPHandler(debuglevel = debuglevel)
             https_handler = urllib.request.HTTPSHandler(context = context, debuglevel = debuglevel)
             error_handler = urllib.request.HTTPErrorProcessor()

             opener = urllib.request.build_opener(cookie_handler, auth_handler, http_handler, https_handler, error_handler)
             urllib.request.install_opener(opener)

             userpass = username + ":" + password
             auth_string = base64.b64encode(userpass.encode()).decode()
             headers = {"Authorization" : "Basic %s" % auth_string}

             version_url = url + "/api/version"
             version_request = urllib.request.Request(version_url, None, headers)
             version_response = urllib.request.urlopen(version_request, timeout = self.timeout_sec).read()

             if version_response:
               self.api_version = version_response.decode()

             hosts_url = url + "/api/" + self.api_version + "/hosts"
             hosts_response = opener.open(hosts_url, timeout = self.timeout_sec).read()
             host_names = {}

             if hosts_response:
                hosts = json.loads(hosts_response.decode())
                for host in hosts["items"]:
                    hostid = host.get("hostId")
                    host_names[hostid] = host.get("hostname")

             clusters_url = url + "/api/" + self.api_version + "/clusters"
             clusters_response = opener.open(clusters_url, timeout = self.timeout_sec).read()

             if clusters_response:
                clusters = json.loads(clusters_response.decode())
                for cluster in clusters["items"]:
                    cluster_name = cluster.get("displayName")
                    cluster_name_updated = cluster_name.replace(" ", "_")
                    hosts = {"hosts" : []}
                    i = 2
                    while cluster_name_updated in self.inventory:
                       if cluster_name_updated + "_" + str(i) in self.inventory:
                          i = i + 1
                       else:
                          cluster_name_updated = cluster_name_updated + "_" + str(i)
                          break
                    self.inventory[cluster_name_updated] = hosts
                    clusters_hosts_url = url + "/api/" + self.api_version + "/clusters/" + cluster_name + "/hosts"
                    clusters_hosts_url = urllib.parse.quote(clusters_hosts_url,":/")
                    clusters_hosts_response = opener.open(clusters_hosts_url, timeout = self.timeout_sec).read()

                    if clusters_hosts_response:
                       hostids = json.loads(clusters_hosts_response.decode())
                       for hostid in hostids["items"]:
                          hostid = hostid.get("hostId")
                          host = host_names[hostid]
                          self.inventory[cluster_name_updated]["hosts"].append(host)

             parsed_url = urlparse(url)
             save_cookie(cookie, self.cookie_path + "." + str(parsed_url.hostname))

        except Exception as e:
             if "Bad credentials" in str(e):
                 if os.path.exists(self.cookie_path):
                     os.remove(self.cookie_path)
                 return 1
             elif "CERTIFICATE_VERIFY_FAILED" in str(e):
                 sys.stderr.write("URL: " + url + "\n")
                 sys.stderr.write(str(e) + "\n")
                 sys.stderr.write("\nThe root CA check for Cloudera Manager server(s) can be disabled or enabled with this:\n\n")
                 sys.stderr.write("export CM_DISABLE_CA_CHECK=True\n\n")
                 sys.exit(-1)
             else:
                 sys.stderr.write("URL: " + url + "\n")
                 sys.stderr.write(str(e) + "\n")
                 sys.exit(-1)

    def read_args(self):
        parser = argparse.ArgumentParser()
        parser.add_argument("--list", help = "List the clusters and the hosts", action = "store_true")
        parser.add_argument("--host", help = "", action = "store")
        parser.add_argument("--refresh-cache", help = "Refresh the (Cloudera Manager) inventory cache" , action = "store_true")
        self.args = parser.parse_args()

        if self.args.refresh_cache:
            self.option_refresh_cache()

        if self.args.list:
            self.option_list()

        if self.args.host:
            self.option_host()

    def option_list(self):
        if os.path.exists(self.json_path) and (time.time() - os.path.getmtime(self.json_path) < self.cache_time_sec):
           json_data = open(self.json_path, "r")
           self.inventory = json.load(json_data)
           print(json.dumps(self.inventory, sort_keys = True, indent = 4, separators=(",", ": ")))
           sys.exit(0)

        self.option_refresh_cache()
        print(json.dumps(self.inventory, sort_keys = True, indent = 4, separators=(",", ": ")))

    def option_refresh_cache(self):
        cm_urls = os.getenv("CM_URL")
        username = os.getenv("CM_USERNAME")
        password = ""

        if cm_urls == None:
           sys.stderr.write("CM_URL shell variable has not been set, configure the Cloudera Manager's URL. Examples:\n\n")
           sys.stderr.write("export CM_URL=https://host-10-17-101-73.coe.example.com:7183\n")
           sys.stderr.write("export CM_URL=https://host-10-17-101-73.coe.example.com:7183,http://nightly512-1.vpc.example.com:7180\n\n")
           sys.exit(-1)

        if username == None:
           sys.stderr.write("CM_USERNAME shell variable has not been set, configure the Cloudera Manager's username. Example:\n\n")
           sys.stderr.write("export CM_USERNAME=username\n\n")
           sys.exit(-1)

        urls = cm_urls.split(",")
       
        for url in urls:
           url = url.strip()
           parsed_url = urlparse(url)
           cookie_path = self.cookie_path + "." + str(parsed_url.hostname)
           if os.path.isfile(cookie_path):
              cookie = load_cookie(cookie_path)
              result = self.cm_connection(url, username, password, cookie)
              if result != 1:
                 continue

           cookie = http.cookiejar.LWPCookieJar()
           while True:
              prompt_message = "Cloudera Manager's URL: " + url + "\nCloudera Manager's password: " 
              password = getpass.getpass(prompt_message)
              result = self.cm_connection(url, username, password, cookie)
              if result != 1:
                 break
              user_input = input("Cloudera Manager's username (default: %s): " % username)
              if user_input:
                 username = user_input.strip()

        with open(self.json_path, "w") as outfile:
           json.dump(self.inventory, outfile)
    
    def option_host(self):
        print(json.dumps(self.inventory, sort_keys = True, indent = 4, separators=(",", ": ")))

if __name__ == "__main__":
    ClouderaManagerInventory()