[Bug/gh-71] We need to make sure cron job starts after django app (#72)

* [bug/gh-71] attempted to fix by pinging website but unable to find solution; temporary fix implemented by waiting a fixed amount of time

* [bug/gh-71] implemented loop that continuously pings eDrop until it responds successfully

* [bug/gh-71] made pinging host and port configurable

* [bug/gh-71] set priorty of gunicorn process ahead of cron job

* [bug-gh-71] modified netcat loop to a curl loop

* [gh-17] added a comment explaining the script change

* [gh-71] we don't need to install netcat

---------

Co-authored-by: Julia Damerow <[email protected]>

Author: jophals

edrop-cron.sh

@@ -15,5 +15,21 @@ export DJANGO_SETTINGS_MODULE=$DJANGO_SETTINGS_MODULE
 export PYTHONPATH=$DJANGODIR:$PYTHONPATH
 mkdir -p /edrop/logs
 
+# When a new container is being created, eDROP Connector takes a long time to startup because all dependencies
+# need to be installed first. The cron job, however, needs all dependencies to be installed before it can start up.
+# Since we know that once eDROP Connector is up and running, all dependencies are installed, we just wait till we
+# get a 200 response from the web app before starting the cron job.
+# Using the autorestart property of Supervisor does not seem to work when deployed in production, hence this work around.
+while true; do 
+    sleep 5;
+    http_response=$(curl -s -o /dev/null -I -w "%{http_code}\n" http://localhost:8000/$APP_ROOT);
+    if [ $http_response == "200" ]; then
+        break
+    else
+        echo "eDROP Connector is not yet responding."
+    fi
+done
+echo "eDROP Connector has started successfully"
+
 # start the cron job
 python manage.py runapscheduler

supervisord.conf

@@ -22,7 +22,7 @@ stopasgroup=true
 
 ; if rabbitmq is supervised, set its priority higher
 ; so it starts first
-priority=998
+priority=997
 
 [program:edrop-cron]
 command=/edrop/edrop-cron.sh