Я использую микросервисную колбу с gunicorn для отправки запросов в службу, размещенную на движке Google Kubernetes. Микросервис докеризован и размещается как модуль в google kubernetes engine. После локального тестирования я развернул его, но получаю ошибку CrashLoopBackOff. Я не получаю никаких ошибок в локальной контейнерной среде. Журналы моего модуля:
[2019-03-15 08:41:13 +0000] [1] [INFO] Starting gunicorn 19.9.0
[2019-03-15 08:41:13 +0000] [1] [INFO] Listening at: http://0.0.0.0:8080 (1)
[2019-03-15 08:41:13 +0000] [1] [INFO] Using worker: threads
[2019-03-15 08:41:13 +0000] [12] [INFO] Booting worker with pid: 12
Failed to find application object 'app' in 'app'
[2019-03-15 08:41:13 +0000] [12] [INFO] Worker exiting (pid: 12)
[2019-03-15 08:41:13 +0000] [13] [INFO] Booting worker with pid: 13
Failed to find application object 'app' in 'app'
[2019-03-15 08:41:13 +0000] [13] [INFO] Worker exiting (pid: 13)
[2019-03-15 08:41:13 +0000] [1] [INFO] Shutting down: Master
[2019-03-15 08:41:13 +0000] [1] [INFO] Reason: App failed to load.
Это похоже на ошибку с gunicorn .
Моя структура папок :
.
├── app.py
├── app.yaml
├── config.py
├── data
│ └── object_detection.pbtxt
├── Dockerfile
├── filename.jpg
├── helper.py
├── object_2.py
├── object_detection
│ ├── core
│ │ ├── anchor_generator.py
│ └── vrd_evaluation_test.py
├── object_detection_grpc_client.py
├── requirements.txt
└── tensorflow_serving
├── apis
└── regression_pb2.py
Код app.py :
import logging
from flask import Flask, request, jsonify
from config import auth_secret_token, PORT, DEBUG_MODE
from helper import check_parameters
from object_detection_grpc_client import main
app = Flask(__name__)
def check_authorization(request):
try:
if not 'Auth-token' in request.headers:
return jsonify({'error': 'unauthorized access'}), 401
token = request.headers['Auth-token']
if token != auth_secret_token:
return jsonify({'error': 'unauthorized access'}), 401
return "ok", 200
except Exception as e:
return jsonify({'error': 'unauthorized access'}), 401
@app.route("/", methods=['POST'])
def hello():
info, status_code = check_authorization(request)
if status_code != 200:
return info, status_code
else:
status, status_code = check_parameters(request.form)
if status_code != 200:
return status, status_code
else:
score = main()
response = {"status": "success", "score": score, "customer_id":(request.form["cust_id"])}
return jsonify(response), status_code
if __name__ == "__main__":
app.run(host="0.0.0.0", port=PORT, debug=DEBUG_MODE)
Код config.py :
from os import environ as env
import multiprocessing
PORT = int(env.get("PORT", 8080))
DEBUG_MODE = int(env.get("DEBUG_MODE", 1))
# Gunicorn config
bind = ":" + str(PORT)
workers = multiprocessing.cpu_count() * 2 + 1
threads = 2 * multiprocessing.cpu_count()
auth_secret_token = "token"
server='A.B.C.D:9000'
model_name="mymodel"
input_image='filename.jpg'
label_map="./data/object_detection.pbtxt"
Dockerfile :
FROM python:3.5.2
RUN apt update
WORKDIR /app
ADD requirements.txt /app/requirements.txt
RUN pip install -r /app/requirements.txt
ADD . /app
ENV PORT 8080
CMD ["gunicorn", "app:app", "--config=config.py"]
Файл развертывания app.yaml :
apiVersion: extensions/v1beta1
kind: Deployment
metadata:
name: mymodel-client
labels:
name: mymodel-client
spec:
replicas: 1
selector:
matchLabels:
name: mymodel-client
template:
metadata:
name: mymodel-client
labels:
name: mymodel-client
spec:
containers:
- name: mymodel-client
image: gcr.io/image-identification/mymodel-client:v1
ports:
- containerPort: 8080
resources:
requests:
memory: 256Mi
limits:
memory: 512Mi
env:
- name: DEBUG_MODE
value: "1"
Что здесь не так?