From ea043a773900a92de3958833dd5ea6e3eacdab5a Mon Sep 17 00:00:00 2001
From: Qi Zhao <zhaoqi99@outlook.com>
Date: Wed, 31 Mar 2021 17:18:28 +0800
Subject: [PATCH] feat: support collect Hadoop that turning on https

---
 collectors/0/hadoop_datanode.py |  2 +-
 collectors/0/hadoop_namenode.py |  2 +-
 collectors/lib/hadoop_http.py   | 23 ++++++++++++++++++++---
 3 files changed, 22 insertions(+), 5 deletions(-)

diff --git a/collectors/0/hadoop_datanode.py b/collectors/0/hadoop_datanode.py
index 45ac73187..f976b4f6f 100755
--- a/collectors/0/hadoop_datanode.py
+++ b/collectors/0/hadoop_datanode.py
@@ -41,7 +41,7 @@ class HadoopDataNode(HadoopHttp):
     """
 
     def __init__(self):
-        super(HadoopDataNode, self).__init__('hadoop', 'datanode', 'localhost', 50075)
+        super(HadoopDataNode, self).__init__('hadoop', 'datanode', 'localhost', 50075, https_port=50475)
 
     def emit(self):
         current_time = int(time.time())
diff --git a/collectors/0/hadoop_namenode.py b/collectors/0/hadoop_namenode.py
index 67ec9fac3..7eeb9335a 100755
--- a/collectors/0/hadoop_namenode.py
+++ b/collectors/0/hadoop_namenode.py
@@ -39,7 +39,7 @@ class HadoopNameNode(HadoopHttp):
     """
 
     def __init__(self):
-        super(HadoopNameNode, self).__init__('hadoop', 'namenode', 'localhost', 50070)
+        super(HadoopNameNode, self).__init__('hadoop', 'namenode', 'localhost', 50070, https_port=50470)
 
     def emit(self):
         current_time = int(time.time())
diff --git a/collectors/lib/hadoop_http.py b/collectors/lib/hadoop_http.py
index e99aee8ad..481538033 100644
--- a/collectors/lib/hadoop_http.py
+++ b/collectors/lib/hadoop_http.py
@@ -12,6 +12,7 @@
 # of the GNU Lesser General Public License along with this program.  If not,
 # see <http://www.gnu.org/licenses/>.
 
+import ssl
 import sys
 
 try:
@@ -26,10 +27,10 @@
 
 try:
     # noinspection PyCompatibility
-    from http.client import HTTPConnection
+    from http.client import HTTPConnection, HTTPSConnection
 except ImportError:
     # noinspection PyUnresolvedReferences,PyCompatibility
-    from httplib import HTTPConnection
+    from httplib import HTTPConnection, HTTPSConnection
 
 
 EXCLUDED_KEYS = (
@@ -52,14 +53,22 @@ def recurse(key,value):
         yield (key,value)
 
 class HadoopHttp(object):
-    def __init__(self, service, daemon, host, port, uri="/jmx"):
+    def __init__(self, service, daemon, host, port, https_port=None, uri="/jmx"):
         self.service = service
         self.daemon = daemon
         self.port = port
+        self.https_port = https_port
         self.host = host
         self.uri = uri
         self.server = HTTPConnection(self.host, self.port)
         self.server.auto_open = True
+        if self.https_port:
+            self.https_server = HTTPSConnection(
+                self.host,
+                self.https_port,
+                context=ssl._create_unverified_context(),
+                timeout=10)
+            self.https_server.auto_open = True
 
     def request(self):
         try:
@@ -67,6 +76,14 @@ def request(self):
             resp = self.server.getresponse().read()
         except:
             resp = '{}'
+            if self.https_port:
+                try:
+                    self.https_server.request('GET', self.uri)
+                    resp = self.https_server.getresponse().read()
+                except:
+                    resp = '{}'
+                finally:
+                    self.https_server.close()
         finally:
             self.server.close()
         return json.loads(resp)