Install API server and Controller

  1. Introduction
  2. Install dependencies
  3. Set up database
  4. Update config.yml
  5. Update nginx configuration
  6. Install arvados-api-server and arvados-controller
  7. Confirm working installation

Introduction

The Arvados core API server consists of four services: PostgreSQL, Arvados Rails API, Arvados Controller, and Nginx.

Here is a simplified diagram showing the relationship between the core services. Client requests arrive at the public-facing Nginx reverse proxy. The request is forwarded to Arvados controller. The controller is able handle some requests itself, the rest are forwarded to the Arvados Rails API. The Rails API server implements the majority of business logic, communicating with the PostgreSQL database to fetch data and make transactional updates. All services are stateless, except the PostgreSQL database. This guide assumes all of these services will be installed on the same node, but it is possible to install these services across multiple nodes.

Install dependencies

  1. Install PostgreSQL
  2. Install Ruby and Bundler
  3. Install nginx
  4. Install Phusion Passenger

Set up database

  1. Start a shell for the postgres user:
    # su postgres
  2. Generate a new database password:
    postgres$ tr -dc 0-9a-zA-Z </dev/urandom | head -c25; echo
    yourgeneratedpassword
    
    Record this. You’ll need it when you set up the Rails server later.
  3. Create a database user with the password you generated:
    postgres$ createuser --encrypted --no-createrole --no-superuser --pwprompt arvados
      Enter password for new role: yourgeneratedpassword
      Enter it again: yourgeneratedpassword
  4. Create a database owned by the new user:
    postgres$ createdb arvados_production -T template0 -E UTF8 -O arvados
  5. Enable the pg_trgm extension
    postgres$ psql arvados_production -c "CREATE EXTENSION IF NOT EXISTS pg_trgm"
  6. Exit the postgres user shell:
    postgres$ exit

Update config.yml

Starting from an empty config.yml file, add the following configuration keys.

Tokens

    SystemRootToken: "$system_root_token"
    ManagementToken: "$management_token"
    Collections:
      BlobSigningKey: "$blob_signing_key"

These secret tokens are used to authenticate messages between Arvados components.

  • SystemRootToken is used by Arvados system services to authenticate as the system (root) user when communicating with the API server.
  • ManagementToken is used to authenticate access to system metrics.
  • Collections.BlobSigningKey is used to control access to Keep blocks.

Each token should be a string of at least 50 alphanumeric characters. You can generate a suitable token with the following command:

~$ tr -dc 0-9a-zA-Z </dev/urandom | head -c50 ; echo

PostgreSQL.Connection

    PostgreSQL:
      Connection:
        host: localhost
        user: arvados
        password: $postgres_password
        dbname: arvados_production

Replace the $postgres_password placeholder with the password you generated during database setup .

Services

    Services:
      Controller:
        ExternalURL: "https://ClusterID.example.com"
        InternalURLs:
          "http://localhost:8003": {}
      RailsAPI:
        # Does not have an ExternalURL
        InternalURLs:
          "http://localhost:8004": {}

Replace ClusterID.example.com with the hostname that you previously selected for the API server.

The Services section of the configuration helps Arvados components contact one another (service discovery). Each service has one or more InternalURLs and an ExternalURL. The InternalURLs describe where the service runs, and how the Nginx reverse proxy will connect to it. The ExternalURL is how external clients contact the service.

Update nginx configuration

Use a text editor to create a new file /etc/nginx/conf.d/arvados-api-and-controller.conf with the following configuration. Options that need attention are marked in red.

proxy_http_version 1.1;

# When Keep clients request a list of Keep services from the API
# server, use the origin IP address to determine if the request came
# from the internal subnet or it is an external client.  This sets the
# $external_client variable which in turn is used to set the
# X-External-Client header.
#
# The API server uses this header to choose whether to respond to a
# "available keep services" request with either a list of internal keep
# servers (0) or with the keepproxy (1).
#
# Following the example here, update the 10.20.30.0/24 netmask
# to match your private subnet.
# Update 1.2.3.4 and add lines as necessary with the public IP
# address of all servers that can also access the private network to
# ensure they are not considered 'external'.

geo $external_client {
  default        1;
  127.0.0.0/24   0;
  10.20.30.0/24  0;
  1.2.3.4/32     0;
}

# This is the port where nginx expects to contact arvados-controller.
upstream controller {
  server     localhost:8003  fail_timeout=10s;
}

server {
  # This configures the public https port that clients will actually connect to,
  # the request is reverse proxied to the upstream 'controller'

  listen       443 ssl;
  server_name  ClusterID.example.com;

  ssl_certificate     /YOUR/PATH/TO/cert.pem;
  ssl_certificate_key /YOUR/PATH/TO/cert.key;

  # Refer to the comment about this setting in the passenger (arvados
  # api server) section of your Nginx configuration.
  client_max_body_size 128m;

  location / {
    proxy_pass            http://controller;
    proxy_redirect        off;
    proxy_connect_timeout 90s;
    proxy_read_timeout    300s;

    proxy_set_header      Host              $http_host;
    proxy_set_header      Upgrade           $http_upgrade;
    proxy_set_header      Connection        "upgrade";
    proxy_set_header      X-External-Client $external_client;
    proxy_set_header      X-Forwarded-For   $proxy_add_x_forwarded_for;
    proxy_set_header      X-Forwarded-Proto https;
    proxy_set_header      X-Real-IP         $remote_addr;
  }
}

server {
  # This configures the Arvados API server.  It is written using Ruby
  # on Rails and uses the Passenger application server.

  listen localhost:8004;
  server_name localhost-api;

  root /var/www/arvados-api/current/public;
  index  index.html index.htm index.php;

  passenger_enabled on;

  # If you are using RVM, uncomment the line below.
  # If you're using system ruby, leave it commented out.
  #passenger_ruby /usr/local/rvm/wrappers/default/ruby;

  # This value effectively limits the size of API objects users can
  # create, especially collections.  If you change this, you should
  # also ensure the following settings match it:
  # * `client_max_body_size` in the previous server section
  # * `API.MaxRequestSize` in config.yml
  client_max_body_size 128m;
}

Install arvados-api-server and arvados-controller

Red Hat and Centos

# yum install arvados-api-server arvados-controller

Debian and Ubuntu

# apt-get install arvados-api-server arvados-controller

Start the service

# systemctl enable --now arvados-controller
# systemctl status arvados-controller
[...]

If systemctl status indicates it is not running, use journalctl to check logs for errors:

# journalctl -n12 --unit arvados-controller

Confirm working installation

Confirm working controller:

$ curl https://ClusterID.example.com/arvados/v1/config

Confirm working Rails API server:

$ curl https://ClusterID.example.com/discovery/v1/apis/arvados/v1/rest

Confirm that you can use the system root token to act as the system root user:


$ curl -H "Authorization: Bearer $system_root_token" https://ClusterID.example.com/arvados/v1/users/current

Troubleshooting

If you are getting TLS errors, make sure the ssl_certificate directive in your nginx configuration has the full certificate chain

Logs can be found in /var/www/arvados-api/current/log/production.log and using journalctl -u arvados-controller.

See also the admin page on Logging .


Previous: Configuration reference Next: Install Keepstore servers

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.