This repository has been archived by the owner on Aug 16, 2020. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 0
/
taskqueue.py
executable file
·212 lines (161 loc) · 5.04 KB
/
taskqueue.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
import sys
import types
import queue
import threading
from time import time
from hashlib import md5
import dill
import redis
import requests
from task import Task
import thriftpy
from thriftpy.rpc import make_server, make_client
dequeue_thrift = thriftpy.load('dequeue.thrift', module_name='dequeue_thrift')
enqueue_thrift = thriftpy.load('enqueue.thrift', module_name='enqueue_thrift')
def get_ip(interface='eth0'):
from netifaces import AF_INET
import netifaces as ni
return ni.ifaddresses(interface)[AF_INET][0]['addr']
redis_ip = "127.0.0.1"
redis_port = 6379
master_ip = "127.0.0.1"
master_port = 9091
rconn = redis.Redis(redis_ip, redis_port)
tq = queue.Queue() # queue of task objs
tasks = {}
workers = []
threads = []
def top_n(tqueue, n):
result = []
iterable = tqueue.__iter__()
for _ in range(n):
try:
result.append(next(iterable))
except StopIteration:
break
return result
def top_pending(n):
return top_n(tq, 5)
def last_running():
result = []
for worker in workers:
result.append(getstr('worker:' + worker.id + '.current'))
none_count = result.count(None)
for _ in none_count:
result.remove(None)
return result
def getint(id):
try:
return int(rconn.get(id).decode('utf-8'))
except AttributeError:
return None
def getstr(id):
try:
return rconn.get(id).decode('utf-8')
except AttributeError:
return None
def send_to_worker(task, ip, port):
client = make_client(dequeue_thrift.RPC, ip, port)
client.task_service(task)
def send_to_taskqueue(task, task_id, ip, port):
client = make_client(enqueue_thrift.RPC, ip, port)
client.task_service(task, task_id)
def fetch_task(task_id):
try:
task = tasks[task_id]
return task
except KeyError:
print("task not found")
return Task(None)
class Producer:
def __init__(self):
self.r = rconn
self.available = False
def set_available(self):
self.available = True
def set_not_available(self):
self.available = False
def enqueue(self, work):
source = work.__code__
task = Task(source)
data = dill.dumps(task)
self.r.set(task.id, 'pending')
send_to_taskqueue(data, task.id, master_ip, master_port)
class Worker:
def __init__(self, ip, port):
self.r = rconn
self.id = md5((ip + str(port)).encode()).hexdigest()[:8]
self.ip = ip
self.port = port
self.set_available()
def set_available(self):
self.r.set('worker:' + self.id + '.available', '1')
def set_not_available(self, task_id):
self.r.set('worker:' + self.id + '.available', '0')
self.r.set('worker:' + self.id + '.current', task_id)
def is_available(self):
state = self.r.get('worker:' + self.id + '.available').decode('utf-8')
if state == '0':
return False
else:
return True
def task_service(self, task):
task = dill.loads(task)
# add_task(task)
self.set_not_available(task.id)
self.work(task)
def work(self, task):
self.r.set(task.id, 'running')
self.r.set('worker:' + self.id + '.current', task.id)
task.running_time = time()
try:
run = types.FunctionType(task.data, globals(), 'run')
task.result = run()
self.r.set(task.id, 'finished')
self.r.incr('worker:' + self.id + ".count_success")
except Exception as error_msg:
print(error_msg, file=sys.stderr)
self.r.set(task.id, 'failed')
self.r.incr('worker:' + self.id + ".count_failed")
self.r.lpush('fail', task.id)
task.running_time = time() - task.running_time
self.set_available()
return task.result
def add_worker(ip, port):
worker = Worker(ip, port)
workers.append(worker)
rconn.lpush("workers", "{} {} {}".format(worker.id, ip, port))
def add_task(task, task_id):
tasks[task_id] = task
def round_robin():
offset = 0
while True:
if workers[offset].is_available():
worker = workers[offset]
task = tq.get()
rconn.incr("output")
send_to_worker(task['data'], worker.ip, worker.port)
tasks[task['id']] = task['data']
offset += 1
offset = offset % len(workers)
class QueueHandler:
def task_service(self, task, task_id):
rconn.incr("input")
tq.put({'data': task, 'id': task_id})
tasks[task_id] = task
def listen():
server = make_server(enqueue_thrift.RPC,
QueueHandler(), master_ip, master_port)
server.serve()
def main():
add_worker('127.0.0.1', 9089)
add_worker('127.0.0.1', 9090)
round_robin()
# Threading
if __name__ == '__main__':
main_thread = threading.Thread(target=main)
threads.append(main_thread)
main_thread.start()
listener_thread = threading.Thread(target=listen)
threads.append(listener_thread)
listener_thread.start()