update

.gitignore

@@ -4,12 +4,15 @@ status.xml
 
 
 
-# IDEA
-
+# IDEA IntelliJ
 *.iml
 .idea/
 
 
+# Tmp files
+*.swp
+*~
+
 
 
 # Byte-compiled / optimized / DLL files

Dockerfile

@@ -0,0 +1,17 @@
+FROM nvidia/cuda:9.0-base
+
+MAINTAINER Newnius <newnius.cn@gmail.com>
+
+RUN apt update && \
+	apt install -y python3 python3-pip
+
+RUN pip3 install docker kafka
+
+ADD bootstrap.sh /etc/bootstrap.sh
+
+ADD yao-agent.py /root/yao-agent.py
+ADD server.py /root/server.py
+
+WORKDIR /root
+
+CMD ["/etc/bootstrap.sh"]

bootstrap.sh

@@ -0,0 +1,6 @@
+#!/usr/bin/env bash
+
+
+python3 /root/yao-agent.py &
+
+python3 /root/server.py

executor.py

@@ -0,0 +1,22 @@
+import docker
+
+
+def run():
+	client = docker.from_env()
+	#print(client.containers.run(image="alpine", command="echo 'Hello World'", environment={"KEY": "value"}))
+	print(client.containers.run(image="nvidia/cuda:9.0-base", command="nvidia-smi", environment={"KEY": "value"}, runtime="nvidia"))
+
+
+def run_in_background():
+	client = docker.from_env()
+	container = client.containers.run("alpine", ["echo", "hello", "world"], detach=True)
+	print(container.id)
+
+
+def list_containers():
+	client = docker.from_env()
+	for container in client.containers.list():
+		print(container.id)
+
+
+run()

server.py

@@ -0,0 +1,108 @@
+#!/usr/bin/python
+from http.server import BaseHTTPRequestHandler, HTTPServer
+import cgi
+import docker
+import json
+from urllib import parse
+
+PORT_NUMBER = 8000
+
+
+# This class will handles any incoming request from
+# the browser
+class MyHandler(BaseHTTPRequestHandler):
+	# Handler for the GET requests
+	def do_GET(self):
+		req = parse.urlparse(self.path)
+		query = parse.parse_qs(req.query)
+
+		if req.path == "/ping":
+			# Open the static file requested and send it
+			self.send_response(200)
+			self.send_header('Content-type', 'application/json')
+			self.end_headers()
+			self.wfile.write(bytes("pong", "utf-8"))
+
+		elif req.path == "/logs":
+			id = query['id'][0]
+			client = docker.from_env()
+			container = client.containers.get(id)
+
+			msg = {'code': 0, 'logs': container.logs().decode()}
+			self.send_response(200)
+			self.send_header('Content-type', 'application/json')
+			self.end_headers()
+			self.wfile.write(bytes(json.dumps(msg), "utf-8"))
+
+		else:
+			self.send_error(404, 'File Not Found: %s' % self.path)
+
+	# Handler for the POST requests
+	def do_POST(self):
+		if self.path == "/create":
+			form = cgi.FieldStorage(
+				fp=self.rfile,
+				headers=self.headers,
+				environ={
+					'REQUEST_METHOD': 'POST',
+					'CONTENT_TYPE': self.headers['Content-Type'],
+				})
+			docker_image = form["image"].value
+			docker_cmd = form["cmd"].value
+
+			print(docker_image)
+			print(docker_cmd)
+
+			client = docker.from_env()
+			container = client.containers.run(
+				image=docker_image,
+				command=docker_cmd,
+				environment={"key": "value"},
+				runtime="nvidia",
+				detach=True
+			)
+
+			msg = {"code": 0, "id": container.id}
+
+			self.send_response(200)
+			self.send_header('Content-type', 'application/json')
+			self.end_headers()
+			self.wfile.write(bytes(json.dumps(msg), "utf-8"))
+
+		elif self.path == "/stop":
+			form = cgi.FieldStorage(
+				fp=self.rfile,
+				headers=self.headers,
+				environ={
+					'REQUEST_METHOD': 'POST',
+					'CONTENT_TYPE': self.headers['Content-Type'],
+				})
+			id = form["id"].value
+
+			client = docker.from_env()
+			container = client.containers.get(id)
+			container.stop()
+			msg = {"code": 0}
+
+			self.send_response(200)
+			self.send_header('Content-type', 'application/json')
+			self.end_headers()
+			self.wfile.write(bytes(json.dumps(msg), "utf-8"))
+		else:
+			self.send_error(404, 'File Not Found: %s' % self.path)
+
+
+try:
+	# Create a web server and define the handler to manage the
+	# incoming request
+	server = HTTPServer(('', PORT_NUMBER), MyHandler)
+	print('Started httpserver on port ', PORT_NUMBER)
+
+	# Wait forever for incoming htto requests
+	server.serve_forever()
+
+except KeyboardInterrupt:
+	print('^C received, shutting down the web server')
+
+
+server.socket.close()

yao-agent.py

@@ -6,7 +6,6 @@ from xml.dom.minidom import parse
 import xml.dom.minidom
 from kafka import KafkaProducer
 
-
 ClientID = os.getenv('ClientID', 1)
 KafkaBrokers = os.getenv('KafkaBrokers', 'localhost:9092').split(',')
 
@@ -18,7 +17,7 @@ def main():
 			status, msg_gpu = execute(['nvidia-smi', '-q', '-x', '-f', 'status.xml'])
 			if not status:
 				print("execute failed, ", msg_gpu)
-				continue
+			continue
 			report_msg()
 			time.sleep(interval)
 		except Exception as e:
@@ -55,6 +54,16 @@ def report_msg():
 			'temperature_gpu': gpu.getElementsByTagName('temperature')[0].getElementsByTagName('gpu_temp')[0].childNodes[0].data,
 			'power_draw': gpu.getElementsByTagName('power_readings')[0].getElementsByTagName('power_draw')[0].childNodes[0].data
 		}
+
+		stat['fan_speed'] = int(float(stat['fan_speed'].split(' ')[0]))
+		stat['memory_total'] = int(float(stat['memory_total'].split(' ')[0]))
+		stat['memory_free'] = int(float(stat['memory_free'].split(' ')[0]))
+		stat['memory_used'] = int(float(stat['memory_used'].split(' ')[0]))
+		stat['utilization_gpu'] = int(float(stat['utilization_gpu'].split(' ')[0]))
+		stat['utilization_mem'] = int(float(stat['utilization_mem'].split(' ')[0]))
+		stat['temperature_gpu'] = int(float(stat['temperature_gpu'].split(' ')[0]))
+		stat['power_draw'] = int(float(stat['power_draw'].split(' ')[0]))
+
 		stats.append(stat)
 
 	post_fields = {'id': ClientID, 'status': stats}