• This repository has been archived on 03/Nov/2020
  • Stars
    star
    106
  • Rank 325,871 (Top 7 %)
  • Language
    Python
  • Created almost 9 years ago
  • Updated about 4 years ago

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

Start a cluster in EC2 for dask.distributed

ARCHIVED

As of November 3rd, 2020 this respository is now archived. Please consult the Dask Cloud docs page for more information on deploying Dask with cloud resources.

Dask EC2 Build Status Coverage Status

Easily launch a cluster on Amazon EC2 configured with dask.distributed, Jupyter Notebooks, and Anaconda.

Installation

You also install dask-ec2 using pip:

$ pip install dask-ec2

You can also install dask-ec2 and its dependencies from the conda-forge repository using conda:

$ conda install dask-ec2 -c conda-forge

Usage

Note: dask-ec2 uses boto3 to interact with Amazon EC2. You can configure your AWS credentials using Environment Variables or Configuration Files.

The dask-ec2 up command can be used to create and provision a cluster on Amazon EC2:

$ dask-ec2 up --help
Usage: dask-ec2 up [OPTIONS]

Options:
  --keyname TEXT                Keyname on EC2 console  [required]
  --keypair PATH                Path to the keypair that matches the keyname
                                [required]
  --name TEXT                   Tag name on EC2
  --tags TEXT                   Additional EC2 tags.  Comma separated K:V
                                pairs: K1:V1,K2:V2
  --region-name TEXT            AWS region  [default: us-east-1]
  --vpc-id TEXT                 EC2 VPC ID
  --subnet-id TEXT              EC2 Subnet ID on the VPC
  --iaminstance-name TEXT       IAM Instance Name
  --ami TEXT                    EC2 AMI  [default: ami-d05e75b8]
  --username TEXT               User to SSH to the AMI  [default: ubuntu]
  --type TEXT                   EC2 Instance Type  [default: m3.2xlarge]
  --count INTEGER               Number of nodes  [default: 4]
  --security-group TEXT         Security Group Name  [default: dask-ec2-default]
  --security-group-id TEXT      Security Group ID (overwrites Security Group
                                Name)
  --volume-type TEXT            Root volume type  [default: gp2]
  --volume-size INTEGER         Root volume size (GB)  [default: 500]
  --file PATH                   File to save the metadata  [default:
                                cluster.yaml]
  --provision / --no-provision  Provision salt on the nodes  [default: True]
  --anaconda / --no-anaconda    Bootstrap anaconda  [default: True]
  --dask / --no-dask            Install Dask.Distributed in the cluster
                                [default: True]
  --notebook / --no-notebook    Start a Jupyter Notebook in the head node
                                [default: True]
  --nprocs INTEGER              Number of processes per worker  [default: 1]
  --source / --no-source        Install Dask/Distributed from git master
                                [default: False]
  -h, --help                    Show this message and exit.

The minimal required arguments for the dask-ec2 up command are:

$ dask-ec2 up --keyname my_aws_key --keypair ~/.ssh/my_aws_key.pem

This will create a cluster.yaml in the directory that it was executed, and this file is required to use the other commands in the CLI.

Once a cluster is running, the dask-ec2 command can be used to create or destroy a cluster, ssh into nodes, or other functions:

$ dask-ec2
Usage: dask-ec2 [OPTIONS] COMMAND [ARGS]...

Options:
  --version   Show the version and exit.
  -h, --help  Show this message and exit.

Commands:
  anaconda          Provision anaconda
  dask-distributed  dask.distributed option
  destroy           Destroy cluster
  notebook          Provision the Jupyter notebook
  provision         Provision salt instances
  ssh               SSH to one of the node. 0-index
  up                Launch instances

More Repositories

1

dask

Parallel computing with task scheduling
Python
12,531
star
2

dask-tutorial

Dask tutorial
Jupyter Notebook
1,832
star
3

distributed

A distributed task scheduler for Dask
Python
1,576
star
4

dask-ml

Scalable Machine Learning with Dask
Python
898
star
5

dask-examples

Easy-to-run example notebooks for Dask
Jupyter Notebook
373
star
6

dask-kubernetes

Native Kubernetes integration for Dask
Python
311
star
7

dask-labextension

JupyterLab extension for Dask
TypeScript
311
star
8

dask-searchcv

dask-searchcv is now part of dask-ml: https://github.com/dask/dask-ml
Python
240
star
9

dask-jobqueue

Deploy Dask on job schedulers like PBS, SLURM, and SGE
Python
234
star
10

dask-docker

Docker images for dask
Jupyter Notebook
231
star
11

dask-image

Distributed image processing
Python
210
star
12

dask-xgboost

Python
162
star
13

cachey

Caching based on computation time and storage space
Python
137
star
14

hdfs3

A wrapper for libhdfs3 to interact with HDFS from Python
Python
136
star
15

dask-gateway

A multi-tenant server for securely deploying and managing Dask clusters.
Python
136
star
16

dask-cloudprovider

Cloud provider cluster managers for Dask. Supports AWS, Google Cloud Azure and more...
Python
134
star
17

partd

Concurrent appendable key-value storage
Python
105
star
18

dask-tensorflow

Python
93
star
19

helm-chart

Helm charts for Dask
YAML
91
star
20

dask-expr

Python
86
star
21

dask-lightgbm

Python
79
star
22

dask-glm

Python
76
star
23

zict

Useful Mutable Mappings
Python
69
star
24

dask-yarn

Deploy dask on YARN clusters
Python
69
star
25

dask-gke

kubernetes setup to bootstrap distributed on google container engine
Python
67
star
26

old-dask-examples

Collection of dask example notebooks
Jupyter Notebook
57
star
27

knit

Deprecated, please use https://github.com/jcrist/skein or https://github.com/dask/dask-yarn instead
Python
53
star
28

dask-mpi

Deploy Dask using MPI4Py
Python
52
star
29

dask-stories

Python
40
star
30

dask-drmaa

Deploy Dask on DRMAA clusters
Python
40
star
31

dask-blog

Dask development blog
HTML
30
star
32

crick

Streaming and approximate algorithms. WIP, use at own risk.
Python
24
star
33

community

For general discussion and community planning. Discussion issues welcome.
20
star
34

mtprof

Thread-aware Python profiler hack
Python
17
star
35

dask-benchmarks

asv benchmarks for dask projects
Python
17
star
36

pandas-streaming

Python
16
star
37

dask-tutorial-infrastructure

Cluster for the Dask Tutorial.
Dockerfile
11
star
38

old-dask-yarn

Deprecated, please use https://github.com/jcrist/skein or https://github.com/dask/dask-yarn instead
Python
7
star
39

governance

The governance process and model for Dask
7
star
40

dask-sphinx-theme

Sphinx theme for Dask documentation
Python
6
star
41

dask-ml-benchmarks

Python
5
star
42

dask.github.io

Dask Website
HTML
5
star
43

scipy-tutorials-2018

5
star
44

design-docs

Experimental repo for proposals of future work
2
star
45

.github

2
star
46

dask-org

General dask resources that aren't code
Jupyter Notebook
2
star
47

marketing

Resources and guidelines for marketing Dask
Python
1
star
48

dask-gateway-helm-repo

Repository holding published dask-gateway helm charts
1
star
49

parquet-integration

Integration tests for various parquet readers and writers
Python
1
star