ephemeral docker workloads with jupyter at pytn

Slide 1

Slide 1 text

Ephemeral Docker Services with IPython, Jupyter, and Cloudpipe

Slide 2

Slide 2 text

Hi!

Slide 3

Slide 3 text

Good morning!

Slide 4

Slide 4 text

Kyle Kelley @rgbkrk

Slide 5

Slide 5 text

No content

Slide 6

Slide 6 text

No content

Slide 7

Slide 7 text

No content

Slide 8

Slide 8 text

No content

Slide 9

Slide 9 text

No content

Slide 10

Slide 10 text

No content

Slide 11

Slide 11 text

No content

Slide 12

Slide 12 text

No content

Slide 13

Slide 13 text

♥ I

Slide 14

Slide 14 text

No content

Slide 15

Slide 15 text

Sharing

Slide 16

Slide 16 text

No content

Slide 17

Slide 17 text

Deployment 㱺

Slide 18

Slide 18 text

No content

Slide 19

Slide 19 text

FROM python:3-onbuild CMD [“python”, “./app.py”]

Slide 20

Slide 20 text

from ﬂask import Flask app = Flask(__name__) ! @app.route('/') def hello_pytn(): return 'Hello PyTennessee!' ! if __name__ == '__main__': app.run(‘0.0.0.0’)

Slide 21

Slide 21 text

$ docker build -t pytn .

Slide 22

Slide 22 text

$ docker build -t pytn . Sending build context to Docker daemon 92.67 kB Sending build context to Docker daemon Step 0 : FROM python:3-onbuild # Executing 3 build triggers Trigger 0, COPY requirements.txt /usr/src/app/ Step 0 : COPY requirements.txt /usr/src/app/ Trigger 1, RUN pip install -r requirements.txt Step 0 : RUN pip install -r requirements.txt ---> Running in dea1149d79ba … Successfully installed Flask Werkzeug Jinja2 itsdangerous markupsafe Cleaning up...

Slide 23

Slide 23 text

$ docker build -t pytn . … Trigger 2, COPY . /usr/src/app Step 0 : COPY . /usr/src/app ---> 6067170ab7fe Removing intermediate container 7ba25a7d8c3b Removing intermediate container dea1149d79ba Removing intermediate container 98f8579edc94 Step 1 : CMD python ./app.py ---> Running in 53031fa4ba6d ---> a7324c447943 Removing intermediate container 53031fa4ba6d Successfully built a7324c447943

Slide 24

Slide 24 text

$ docker build -t pytn .

Slide 25

Slide 25 text

$ docker run -p 5000:5000 pytn * Running on http://*:5000/ (Press CTRL+C to quit)

Slide 26

Slide 26 text

No content

Slide 27

Slide 27 text

$ docker run ipython/scipystack

Slide 28

Slide 28 text

ipython/ipython ipython/scipystack ipython/notebook ipython/scipyserver

Slide 29

Slide 29 text

No content

Slide 30

Slide 30 text

No content

Slide 31

Slide 31 text

No content

Slide 32

Slide 32 text

No content

Slide 33

Slide 33 text

“Could we do a live demo for Nature?”

Slide 34

Slide 34 text

“Yes?”

Slide 35

Slide 35 text

No content

Slide 36

Slide 36 text

JUPYTERHUB IN ACTION Kyle visits /user/rgbkrk/ to reach his notebook server

Slide 37

Slide 37 text

CONFIGURABLE HTTP PROXY • Based on node-http-proxy • Proxy routes to localhost, external IPs, etc. • /user/123/ → 127.0.0.1:56790 • Provides an administrative API for adding routes • POST /api/routes/[:path]

Slide 38

Slide 38 text

github.com/jupyter/ conﬁgurable-http-proxy

Slide 39

Slide 39 text

No content

Slide 40

Slide 40 text

No content

Slide 41

Slide 41 text

No content

Slide 42

Slide 42 text

“Yes!”

Slide 43

Slide 43 text

github.com/jupyter/tmpnb

Slide 44

Slide 44 text

/user/reqz/ /user/asdf/ /user/uiop/ :59678 :59632 :59680 Conﬁgurable HTTP Proxy

Slide 45

Slide 45 text

No content

Slide 46

Slide 46 text

tmpnb.org

Slide 47

Slide 47 text

bit.ly/inature

Slide 48

Slide 48 text

> 20,000 notebook servers

Slide 49

Slide 49 text

No content

Slide 50

Slide 50 text

No content

Slide 51

Slide 51 text

No content

Slide 52

Slide 52 text

No content

Slide 53

Slide 53 text

No content

Slide 54

Slide 54 text

What else is ephemeral?

Slide 55

Slide 55 text

compute jobs

Slide 56

Slide 56 text

No content

Slide 57

Slide 57 text

>>> import cloud >>> def add(x, y): ... return x+y >>> add(1, 2) 3 >>> jid = cloud.call(add, ... 1, 2)

Slide 58

Slide 58 text

! >>> cloud.result(jid) 3

Slide 59

Slide 59 text

>>> import cloud >>> def scrape(x,y): ... # Expensive operation >>> xs = [] # list of x’s >>> ys = [] # list of y’s >>> jids = cloud.map(scrape, ... xs, ys)

Slide 60

Slide 60 text

>>> cloud.result(jids)

Slide 61

Slide 61 text

PICLOUD • Chuck a python pickle to the cloud (f, args, kwargs) = pickle.load(sys.stdin) res = f(*args, **kwargs) ! cloudpickle.dump(res, open("/tmp/.result", "wb"))

Slide 62

Slide 62 text

Acquisition

Slide 63

Slide 63 text

No content

Slide 64

Slide 64 text

Slide 65

Slide 65 text

No content

Slide 66

Slide 66 text

Slide 67

Slide 67 text

github.com/cloudpipe

Slide 68

Slide 68 text

No content

Slide 69

Slide 69 text

import multyvac ! # Connect to the cluster multyvac.config.set_key( api_key=rackspace_user, api_secret_key=rackspace_api_key, api_url=‘https://api.cloudpi.pe/v1’) ! ! ! !

Slide 70

Slide 70 text

CLOUDPIPE • Docker Backed • Processes run as non-root • API compatible with multyvac

Slide 71

Slide 71 text

CLOUDPIPE • Open Source • Want Need Python 3 binding, new cloud pickler • Node library in the works

Slide 72

Slide 72 text

Where else can we take things?

Slide 73

Slide 73 text

On Demand IPython Kernels ! github.com/rgbkrk/ kernels-service

Slide 74

Slide 74 text

bit.ly/oreillyjupyter

Slide 75

Slide 75 text

FRONTIER • Put kernels close to data stores, spin out compute • IPython parallel on demand • Compute jobs on demand • Trigger jobs by events