Arvados-in-a-box

Arvbox is a Docker-based self-contained development, demonstration and testing environment for Arvados. It is not intended for production use.

Quick start

$ git clone https://github.com/curoverse/arvados.git
$ cd arvados/tools/arvbox/bin
$ ./arvbox start localdemo

Requirements

  • Linux 3.x+ and Docker 1.9+
  • Minimum of 3 GiB of RAM + additional memory to run jobs
  • Minimum of 3 GiB of disk + storage for actual data

Usage

$ arvbox
Arvados-in-a-box                      http://arvados.org

build   <config>      build arvbox Docker image
rebuild <config>      build arvbox Docker image, no layer cache
start|run <config>  start arvbox container
open       open arvbox workbench in a web browser
shell      enter arvbox shell
ip         print arvbox docker container ip address
host       print arvbox published host
status     print some information about current arvbox
stop       stop arvbox container
restart <config>  stop, then run again
reboot  <config>  stop, build arvbox Docker image, run
reset      delete arvbox arvados data (be careful!)
destroy    delete all arvbox code and data (be careful!)
log <service> tail log of specified service
ls <options>  list directories inside arvbox
cat <files>   get contents of files inside arvbox
pipe       run a bash script piped in from stdin
sv <start|stop|restart> <service> change state of service inside arvbox
clone <from> <to>   clone an arvbox

Configs

dev

Development configuration. Boots a complete Arvados environment inside the container. The “arvados”, “arvado-dev” and “sso-devise-omniauth-provider” code directories along data directories “postgres”, “var”, “passenger” and “gems” are bind mounted from the host file system for easy access and persistence across container rebuilds. Services are bound to the Docker container’s network IP address and can only be accessed on the local host.

In “dev” mode, you can override the default autogenerated settings of Rails projects by adding “application.yml.override” to any Rails project (sso, api, workbench). This can be used to test out API server settings or point Workbench at an alternate API server.

localdemo

Demo configuration. Boots a complete Arvados environment inside the container. Unlike the development configuration, code directories are included in the demo image, and data directories are stored in a separate data volume container. Services are bound to the Docker container’s network IP address and can only be accessed on the local host.

test

Run the test suite.

publicdev

Publicly accessible development configuration. Similar to ‘dev’ except that service ports are published to the host’s IP address and can accessed by anyone who can connect to the host system. See below for more information. WARNING! The public arvbox configuration is NOT SECURE and must not be placed on a public IP address or used for production work.

publicdemo

Publicly accessible development configuration. Similar to ‘localdemo’ except that service ports are published to the host’s IP address and can accessed by anyone who can connect to the host system. See below for more information. WARNING! The public arvbox configuration is NOT SECURE and must not be placed on a public IP address or used for production work.

Environment variables

ARVBOX_DOCKER

The location of Dockerfile.base and associated files used by “arvbox build”.
default: result of $(readlink -f $(dirname $0)/../lib/arvbox/docker)

ARVBOX_CONTAINER

The name of the Docker container to manipulate.
default: arvbox

ARVBOX_BASE

The base directory to store persistent data for arvbox containers.
default: $HOME/.arvbox

ARVBOX_DATA

The base directory to store persistent data for the current container.
default: $ARVBOX_BASE/$ARVBOX_CONTAINER

ARVADOS_ROOT

The root directory of the Arvados source tree
default: $ARVBOX_DATA/arvados

ARVADOS_DEV_ROOT

The root directory of the Arvados-dev source tree
default: $ARVBOX_DATA/arvados-dev

SSO_ROOT

The root directory of the SSO source tree
default: $ARVBOX_DATA/sso-devise-omniauth-provider

ARVBOX_PUBLISH_IP

The IP address on which to publish services when running in public configuration. Overrides default detection of the host’s IP address.

Using Arvbox for Arvados development

The Arvbox section of Hacking Arvados has information about using Arvbox for Arvados development.

Making Arvbox accessible from other hosts

In “dev” and “localdemo” mode, Arvbox can only be accessed on the same host it is running. To publish Arvbox service ports to the host’s service ports and advertise the host’s IP address for services, use publicdev or publicdemo:

$ arvbox start publicdemo

This attempts to auto-detect the correct IP address to use by taking the IP address of the default route device. If the auto-detection is wrong, you want to publish a hostname instead of a raw address, or you need to access it through a different device (such as a router or firewall), set ARVBOX_PUBLISH_IP to the desire hostname or IP address.

$ export ARVBOX_PUBLISH_IP=example.com
$ arvbox start publicdemo

Note: this expects to bind the host’s port 80 (http) for workbench, so you cannot have a conflicting web server already running on the host. It does not attempt to take bind the host’s port 22 (ssh), as a result the arvbox ssh port is not published.

Notes

Services are designed to install and auto-configure on start or restart. For example, the service script for keepstore always compiles keepstore from source and registers the daemon with the API server.

Services are run with process supervision, so a service which exits will be restarted. Dependencies between services are handled by repeatedly trying and failing the service script until dependencies are fulfilled (by other service scripts) enabling the service script to complete.


Previous: Installation overview Next: Prerequisites

The content of this documentation is licensed under the Creative Commons Attribution-Share Alike 3.0 United States licence.
Code samples in this documentation are licensed under the Apache License, Version 2.0.