Celery -Bound Tasks

A task being bound means the first argument to the task will always be the task instance (self).
Bound tasks are needed for retries (using app.Task.retry()), for accessing information about the current task request, and for any additional functionality you add to custom task base classes. Ref

An example to get the tasks ID of current tasks has been added below :

def get_ID(self):
print “hooo”
print self.request.id

It is always a good practice to give names to tasks otherwise it will go for automatic  naming which in some situation may lead to tasks unregistered error.



Celery in Production – Supervisor

In this tutorial, we are going to see how celery is set up in production environment, where both workers and other processes such as monitoring tool like flower are to be run continuously. During development stage, both worker and flower processes used to get stopped somehow forcing me to restart the every now and then. A solution for this, as suggested in official site of Celery, is to make use of tools like Supervisor.

In production you will want to run the worker in the background as a daemon and some times there may be a chance of stopping of celery worker automatically then it should be restarted automatically. To do these tasks you need to use the tools provided like supervisord.

Installing Supervisor

First we need to set up python virtual environment. Then run the following command to create a virtual environment for our demo projects :
virtualenv env

Now move to this folder env and activate the this virtual environment:
source bin/activate

(Now we need to install both celery and rabitMQ in this virtual environment using pip,)

Now install supervisor using the following command :

pip install supervisor

This would create a configuration file named echo_supervisord_conf

Now run the following command to generate the config file :
echo_supervisord_conf > supervisord.conf

This would generate a config file, supervisord.conf where lies all the keys for our magic………….
Now move this file to the destination folder where we have written codes for celery. In my case I have a folder named project inside this env folder (which contains files such as tasks.py etc )

Now cd to projects folder.

Now open the file we have just copied, and add the following lines

command=celery worker -A tasks –loglevel=INFO

command=celery flower -A tasks

Since we need to run both worker and flower processes, they need to be added as two separate tasks as written above. Also we can set these as a group so that they would be started and stopped together. Most of the fields in these are selfexplanatory, however, if you would like to get a clear picture, you can this

Now Starting the demons :
Just go to the project directory (folder where we copied the config file) and open the terminal
and run the following command,

This would start both flower and celery worker as demons.

Stopping Supervisord

If we want stop supervisord, just runthe following command,

killall supervisord

Ref :

HTTP Gateway : Invoking Celery Tasks from Java (Non Python Application) – Part #3

HTTP Gateway is the ideal way for celery tasks invocation and status polling from any non python languages like java. For this, first we have to set up a django application which would handle all the incoming requests for tasks invocation and status polling. As a first step, we need to install python virtual enviroment named env (not mandatory, But I prefer this as this wont affect python packages installed in the machine).

Step1 :
activate the virtual environment and install all required libraries like django,djecelery etc.

Step 2:
copy and paste the django app for http-gateway in this activated virtual environment,env.(I have uploaded the code for this app in github, https://github.com/nishijayaraj/celery-HTTPgateway)

This app contains all tasks in tasks.py file. Now run the app from this activated virtual environment using the following command :
python manage.py runserver

Now with this, our djnago app would start at localhost:8080, to which we can send rest api calls. It is likely to be getting some module not found errors at this point, in case u get them, just try to install the missing packages and libraries in this activated virtual environment. and then run the app again.

The structure of the entire app is described in the image added below :
Screenshot from 2016-07-14 18:21:27

Now we need workers to execute the tasks. So open a terminal and run the following command from the root directory of our application (env/celery-HTTPgateway),

celery worker -A tasks –loglevel=INFO

It would then list all registered tasks there. We would be able to see to tasks,

Now let’s see how we can invoke a celery task and poll its status using REST apis. These rest apis can be called from any programming languages using appropriate native apis.

For the sake of simplicity, here I am using linux curl command for simulating REST api calls.
( / at end of the url is mandatory for this command to work)

Open a terminal and run the following command,

curl -X GET http://localhost:8000/apply/tasks.hello_world/

Then you would a json data as response,

{“ok”: “true”, “task_id”: “0fc2150e-b321-4cc6-aaef-b1ce9b30e7fe”}

The respose contains the id of the invoked tasks which can be used to track its status.

Status Polling:

curl -X GET http://localhost:8000/0fc2150e-b321-4cc6-aaef-b1ce9b30e7fe/status/

{“task”: {“status”: “SUCCESS”, “result”: “Hello world………”, “id”: “0fc2150e-b321-4cc6-aaef-b1ce9b30e7fe”}}

These api would display custom states,something at which celery flower apis fails most of the time. A task named UploadTask has been written in tasks.py file with a view to showcase this feature. For this, first we need to invoke the tasks and then track the status.

Task invocation :

curl -X GET http://localhost:8000/apply/tasks.UploadTask/
Response –
“ok”: “true”,
“task_id”: “cc51e093-372f-42c1-8344-c1def70c544a”

The status checking of above task, can be done :

curl -X GET http://localhost:8000/cc51e093-372f-42c1-8344-c1def70c544a/status/
“task”: {
“status”: “PROGRESS”,
“result”: {
“progress”: 0
“id”: “cc51e093-372f-42c1-8344-c1def70c544a”

References :

Invoking Celery Tasks from Java Application – Part #2

In the previous post we have seen how to invoke a celery tasks from java application. but it was based on sending messge to  rabbitMQ queue using respective rabbitMQ libraries. But in this post, let’s be be familiar with more convenient way or rather using Rest APIs.

For this, we need to install a celery monitoring tool called flower. Not all version of flower is supposed to serve our purpose. What worked for me is the development version. (the command to install is written below)
pip install https://github.com/mher/flower/zipball/master#egg=flower

So let me assume that we have tasks.py with a task named add

def add(x, y):
print x+y

Now run the worker
celery -A tasks worker –loglevel=info

Starting flower
Finally it is time to start flower so that we access/control both tasks and workers using flower REST apis. For that we need to run the following command :

celery flower -A appname (celery flower -A tasks)

Care should be taken to specify the project name in the above command(here tasks) when we start flower because the apis would not work properly otherwise.

Now this can be viewed from the url http://localhost:5555 (or using respective hostname). This has got different tabs to show the status of tasks, workers and so on. So basically what we are going to do is, use the the apis which flower is using for aforementioned feature, directly in our application.

In order to simulate REST api call, throughout this post I am using curl command as I am coming from linux background. This apis can be integrated from any programming languages.

1. Invoking a celery task

curl -X POST -d ‘{“args”:[1,2]}’ http://localhost:5555/api/task/async-apply/tasks.add

this would trigger celery task add with parameters 1 and 2 and would generate an output similar to the following:

“task-id”: “81775ebb-7d88-4e91-b580-b3a2d79fe668”,
“state”: “PENDING”

So this api would return the task id of the generaed task, which can be used for tracking it whenever we want.

2. Retrieving information regarding a specific task using its id

curl -X GET http://localhost:5555/api/task/info/81775ebb-7d88-4e91-b580-b3a2d79fe668

output :
“task-id”: “81775ebb-7d88-4e91-b580-b3a2d79fe668”,
“result”: “‘None'”,
“clock”: 371,
“routing_key”: null,
“retries”: 0,
“failed”: false,
“state”: “SUCCESS”,
“kwargs”: “{}”,
“sent”: false,
“expires”: null,
“exchange”: null,
“started”: 1466248131.745754,
“timestamp”: 1466248131.837694,
“args”: “[1, 2]”,
“worker”: “celery@space-Vostro-3800”,
“revoked”: false,
“received”: 1466248131.744577,
“exception”: null,
“name”: “tasks.add”,
“succeeded”: 1466248131.837694,
“traceback”: null,
“eta”: null,
“retried”: false,
“runtime”: 0.09263942600227892

3. Listing all the tasks sent to workers

curl -X GET http://localhost:5555/api/tasks

output :
“81775ebb-7d88-4e91-b580-b3a2d79fe668”: {
“received”: 1466248131.744577,
“revoked”: false,
“name”: “tasks.add”,
“succeeded”: 1466248131.837694,
“clock”: 371,
“started”: 1466248131.745754,
“timestamp”: 1466248131.837694,
“args”: “[1, 2]”,
“retries”: 0,
“failed”: false,
“state”: “SUCCESS”,
“result”: “‘None'”,
“retried”: false,
“kwargs”: “{}”,
“runtime”: 0.09263942600227892,
“sent”: false,
“uuid”: “81775ebb-7d88-4e91-b580-b3a2d79fe668”
“50c589e1-b613-496f-af1e-c94c04b163dc”: {
“received”: 1466248086.289584,
“revoked”: false,
“name”: “tasks.add”,
“succeeded”: 1466248086.339701,
“clock”: 313,
“started”: 1466248086.291148,
“timestamp”: 1466248086.339701,
“args”: “[4, 3]”,
“retries”: 0,
“failed”: false,
“state”: “SUCCESS”,
“result”: “‘None'”,
“retried”: false,
“kwargs”: “{}”,
“runtime”: 0.049509562999446644,
“sent”: false,
“uuid”: “50c589e1-b613-496f-af1e-c94c04b163dc”

4. Terminating a task
curl -X POST -d ‘terminate=True’ http://localhost:5555/api/task/revoke/81775ebb-7d88-4e91-b580-b3a2d79fe668

References :




Reading Java property file in Python

Accessing  a java property file in a python code is an easy task. For this we need to install, a python module called pyjavaproperties. (There are many other ways in which we can do this. I prefer this module)

For installing this, please run the following command :

sudo pip install http://pypi.python.org/packages/source/p/pyjavaproperties/pyjavaproperties-0.6.tar.gz

How to use it 

we have a property file named  config.properties and which is as follows,

Now open a python ide and add the following lines
from pyjavaproperties import Properties
p = Properties()
p.list()     #will all the properties and its valuesprint                                               print p['user']   #prints Cruchify
Ref :                                                                                                   https://pypi.python.org/pypi/pyjavaproperties                                                           https://www.versioneye.com/python/pyjavaproperties/0.6