Skip to content

About

Local copy of erddap docker compose installation for SMHI erddap server

Resources

Stars

0 stars

Watchers

0 watching

Forks

 
 

Latest commit

 

History

13 Commits

Folders and files

Repository files navigation

ERDDAP Docker

This repository contains the complete toolchain for the ERDDAP server hosted on https://erddap.nodc.se. This includes:

  • Setting up the server environment (with Ansible).
  • A docker-compose solution for the ERDDAP server.
  • A CLI for handling data

Installation

Prerequisites

The intended target for all instructions and scripts is a server running Ubuntu 22.04.4 LTS (Jammy Jellyfish).

Cloning erddap-docker

Using the root user, clone this repository on the server as /srv/erddap/.

Links:

Running the setup

The script scripts/manage.sh setup installs Ansible (with required community collections) and then runs the playbook in ansible/. The playbook does the following:

  • Creates the admin user erddap and makes sure that it can't be logged in using password.
  • Enables the ufw firewall, disabling all incoming traffic except ssh, http, and https.
  • Installs docker and docker-compose.
  • Installs a systemd service that makes sure that the docker-compose file in /docker/ is always running.

At any point, the setup can be verified with the script scripts/manage.sh sync (use flag --check for a read-only dry run).

Creating additional users

You can optionally create personal users that are allowed to add and manage data. This is done with:

bash scripts/manage.sh add-user <username>

The created user will get the correct groups to manage data and also an installation of erdap-cli (see below).

Note that the created user will not be able to log in using a password. The intended method is to add a public SSH key to the file /home/USERNAME/.ssh/authorized_keys. For an introduction to SSH keys, see section Intro to using SSH keys.

Updating the configuration

In docker/docker-compose.yml you should manually change the value for variables starting with "ERDDAP_".

Especially ERDDAP_flagKeyKey should be changed to any string value (the documentation recommends to use a phrase). This value is secret, but you will not have to enter it somewhere else and it can be changed anytime.

Using git when testing out configurations

When experimenting with configurations, there is always a risk that you don't remember what you have changed. Make use of the fact that all files are within git repositories.

See which files and lines have been updated in a repository.

$ git diff

Undo all local changes of a repository:

$ git reset --hard HEAD

Working with data

Default ERDDAP workflow

Out oud the box, ERDDAP handles data in the following way:

  • A single XML file called datasets.xml is the source for all the datasets and how different columns etc. are interpreted.
  • An XML element describing a specific dataset can be generated with the script GenerateDatasetsXml.sh. This is intended as a starting point.
  • Whenever changes have been made to datasets.xml the server must either be restarted or a signal must be sent to the server for each dataset that is to be reloaded. Signaling is made by creating an empty file named after the dataset id inside /erddapData/hardFlag/.

The ERDDAP version used by this project has an additional experimental feature to make it easier to work with datasets. A directory called /datasets.d hosts individual XML files for each dataset. When the server is restarted or the script /init.d/50-datasets.d.sh is called, all XML files are added to datasets.xml. Reloading of datasets must still be performed.

erddapcli

This project includes erddapcli to simplify the above workflow. The tool handles the following:

  • Generate XML for a specific dataset and put it in /datasets.d.
  • Compile all XML files in /datasets.d and trigger update for each dataset.
  • List all datasets described by datasets.xml and in files in /datasets.d.
  • Change the attribute "active" for a specific dataset and trigger update.

Intro to using SSH keys

SSH keys are a convenient way to authenticate to a server without having to enter a password. An SSH key is actually a pair, one public and one private key. The private key should never be shared with anyone, while the public key can be freely shared. Any server that has your public key associated with a user will accept authentication where you use your matching private key.

Generating a key pair

Open a terminal (PowerShell on Windows) and run:

ssh-keygen -t ed25519

Press Enter to accept the default location. Optionally set a passphrase.

Finding the public key locally

The key pair is typically created in a directory named .ssh inside your home directory. For example:

  • /home/<username>/.ssh/ (Linux)
  • /Users/<username>/.ssh/ (macOS)
  • C:\Users\<username>\.ssh\ (Windows)

The public key ends with .pub.

Linux and macOS

Show the public key (from your home directory):

cat .ssh/id_ed25519.pub

Windows

Show the public key (from your home directory):

type .ssh\id_ed25519.pub

Add the public key to the server

On the server, inside the desired user's home directory, add the public key to the file .ssh/authorized_keys. The key should be added on a new line. The username on the server doesn't have to match the username on your local computer.

Connecting to the server using the SSH key

Once the public key has been added to the server, connect with:

ssh <username>@erddap.nodc.se

As long as the private key is stored in the correct location, the key will automatically be used. If you have multiple keys, they will be tested one after another until one is accepted.

If you store your key in a custom location or if you have keys for different users on the server, you can specify which one to use with the -i flag:

ssh <username>@erddap.nodc.se -i my_erddap_key

About

Local copy of erddap docker compose installation for SMHI erddap server

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages