Graceful shutdown implemented and uwsgi removed

developer_docs/docker-image.md

@@ -3,7 +3,7 @@
 The Docker image for this driver includes the following features:
 
 - Installs the driver from a `whl` file created with standard Python setuptools
-- Runs the `ovd` command to start the driver application with a uwsgi container (standard for Python production applications)
+- Runs the `gunicorn --workers $NUM_PROCESSES --bind :$DRIVER_PORT $SSL "osvimdriver:create_wsgi_app()"` command to start the driver application with a Gunicorn based container (standard for Python production applications)
 - Supports installing a development version of Ignition from a `whl` file
 - Supports configuring the uWSGI container implementation used at both build and runtime (also includes configuring the number of processes and threads used by uWSGI container)
 

developer_docs/helm-chart.md

@@ -10,7 +10,6 @@ The Helm chart for this driver includes the following features:
     - Number of replicas of the driver
     - Property groups configured on your application (i.e. override the properties in the default_config file)
     - Log level
-    - uWSGI container used by your application
     - Pod affinity/anti-affinity
     - Node affinity/anti/affinity
     - NodePort

developer_docs/python-app.md

@@ -2,9 +2,9 @@
 
 The Python package in this driver provides a workable application running with Connexion (on top of Flask). The Ignition framework takes care of configuring the application with the APIs and Python objects to handle the requests.
 
-This application is then wrapped by a uWSGI to provide a production ready deployment. 
+This application is then wrapped by Gunicorn to provide a production ready deployment. 
 
-Bin scripts are provided in the `osvimdriver` package to start the driver in development or production mode.
+Gunicorn command is used to run application after deployment, it creates a central master process and worker processes which handle the application.
 
 ## Testing 
 
@@ -28,7 +28,6 @@ This file defines the metadata of the Python package to be built, including any
 
 This file also specifies the entry points to the application, so a user may run the driver on the command line after installation:
     - `ovd-dev` for a development server
-    - `ovd` (`ovd-gunicorn`/`ovd-uswgi`) for a production server
 
 To build a distributable package of your application you will need the `setuptools` and `wheel` Python modules:
 
@@ -48,7 +47,7 @@ This whl file can now be used to install your application with Python:
 python3 -m pip install <path to whl>
 ```
 
-The commands `ovd-dev`, `ovd`, `ovd-gunicorn` and `ovd-uswgi` will now be avaiable from the command line.
+The command `ovd-dev` will now be avaiable from the command line.
 
 ## Configuration
 
@@ -59,4 +58,4 @@ Ignition loads configuration properties from any sources provided to the applica
 - `/var/ovd/ovd_config.yml` - this configuration file is only used in the Helm chart installation. Ignition will search for a configuration file at this path (ignored if not found)
 - `OVD_CONFIG` - set this environment variable to a file path and Ignition will load the configuration file (ignored if the environment variable is not set)
 
-This allows the user flexibility in how to configure the application. When running with Python (using `ovd-dev` or `ovd`) the best approach is to create a `ovd_config.yml` file in the current directory or configure `OVD_CONFIG` with a file path. 
+This allows the user flexibility in how to configure the application. When running with Python (using `ovd-dev` or `gunicorn --workers $NUM_PROCESSES --bind :$DRIVER_PORT $SSL "osvimdriver:create_wsgi_app()"`) the best approach is to create a `ovd_config.yml` file in the current directory or configure `OVD_CONFIG` with a file path. 

docker/Dockerfile

@@ -4,10 +4,6 @@ ARG DRIVER_PORT
 ENV DRIVER_PORT ${DRIVER_PORT:-8292}
 ARG NUM_PROCESSES
 ENV NUM_PROCESSES ${NUM_PROCESSES:-4}
-ARG NUM_THREADS
-ENV NUM_THREADS ${NUM_THREADS:-2}
-ARG WSGI_CONTAINER
-ENV WSGI_CONTAINER ${WSGI_CONTAINER:-gunicorn}
 
 COPY whls/*.whl /whls/
 
@@ -33,4 +29,5 @@ WORKDIR /home/ovd
 
 EXPOSE 8292
 
-CMD ["ovd"]
+CMD if [ $SSL_ENABLED | tr [:upper:] [:lower:] == "true" ]; then SSL="--certfile /var/ovd/certs/tls.crt --keyfile /var/ovd/certs/tls.key" ; fi \
+&& gunicorn --workers $NUM_PROCESSES --bind :$DRIVER_PORT $SSL "osvimdriver:create_wsgi_app()"

helm/os-vim-driver/values.yaml

@@ -39,12 +39,8 @@ app:
       ## as top level fields in the Elasticsearch index.
       LOG_TYPE: logstash
 
-      ## configuration for WSGI container
-      ## WSGI_CONTAINER can be uwsgi or gunicorn
-      WSGI_CONTAINER: gunicorn
       ## the number of processes and threads to spawn to handle requests
       NUM_PROCESSES: "4"
-      NUM_THREADS: "4"
 
     ## ovd_config.yml (driver configuration) overrides
     override:

setup.py

@@ -27,12 +27,10 @@
         'python-novaclient>=13.0.0,<14.0.0',
         'tosca-parser @ git+https://github.com/IBM/tosca-parser.git@accanto',
         'heat-translator @ git+https://github.com/IBM/heat-translator.git@accanto-nfv',
-        'uwsgi==2.0.19.1',
         'gunicorn==20.1.0'
     ],
     entry_points='''
         [console_scripts]
         ovd-dev=osvimdriver.__main__:main
-    ''',
-    scripts=['osvimdriver/bin/ovd-uwsgi', 'osvimdriver/bin/ovd-gunicorn', 'osvimdriver/bin/ovd']
+    '''
 )