README.md

Eremetic synchronous client

A Python client for Eremetic, an Apache Mesos framework to run one-off jobs in Docker containters.

This client simply provides a synchronous interface over the HTTP API.

The major and minor version of this package mirrors the version of Eremetic with which it's compatible. The patch version may vary independently.

Installation

pip install eremetic-asynchronous-client

Example

Preparing the root logger

import sys, logging
logging.root.handlers = []
logging.basicConfig(stream=sys.stdout, level=logging.INFO, format='%(levelname)s - %(message)s')

Preparing (and showing) the request

import eremetic_synchronous_client
request = eremetic_synchronous_client.Request(cpu=1, mem=1024, image='busybox', command='echo $(date)')
request.payload()

Return value

{'network': 'HOST', 'mem': 1024, 'image': 'busybox', 'force_pull_image': False, 'command': 'echo $(date)', 'cpu': 1}

Shooting the request

request.to('http://eremetic-url')

Output

INFO - Task "eremetic-task.232e3359-dead-babe-beef-9872ed82ba90" status progression: QUEUED -> STAGING -> RUNNING -> FINISHED
INFO - TASK_FINISHED: "eremetic-task.232e3359-dead-babe-beef-9872ed82ba90", status page: http://eremetic-url/task/eremetic-task.232e3359-dead-babe-beef-9872ed82ba90

Return value

(u'eremetic-task.232e3359-dead-babe-beef-9872ed82ba90', u'TASK_FINISHED')

Separate submission and task tracking

If you wish to persist the id of the task before blocking your app to wait for it to finish, you can separately submit the task:

task_id = request.submit('http://eremetic-url')

Now you can start tracking the existing task with:

request.track('http://eremetic-url', task_id)

Using this method you can connect to a running task to wait for it to finish or to fetch its result if the task has completed.