Skip to content

INSTALL GALAXY WITH ANSIBLE

Ansible

Ansible

Ansible is an automation system that automates configuration management and application deployment.

Ansible reads instructions (tasks) from a playbook and performs the indicated tasks on target machines (referred to as Hosts), through an ssh connection.

Basically, everything an "administrator" can do using command lines with linux OS (or Mac OS), can be automated with ansible commands that "wraps" these command lines. The power of Ansible (and similar orchestration software, ie Puppet, Chief, etc.) comes from the abstraction of complex suite of commands in the Ansible syntax. Moreover, automation allows to reproduce exactly the desired configuration. Finally, Ansible is idempotent: whatever the initial configuration, it brings the target to the exact same final state. This is useful to repair a broken configuration.

Ansible playbook - GalaxyKickStart

The Ansible "language" (Striclty speaking, Ansible language is not a programming language) is structured. Thus a playbook is not necessarily a single flat file. Multiple tasks can be gathered in a file, a "role" is the execution of a set of tasks, and a playbook can execute multiple roles.

GalaxyKickStart is an Ansible playbook that will

  • install basic dependencies needed for Galaxy
  • Create and manage all the linux users involved in the deployment of Galaxy
  • Install and configure the services required for Galaxy:
    • postgresql (database engine)
    • nginx (web server)
    • docker (containers)
    • proftpd (ftp server)
    • slurm (job manager)
    • supervisor (service manager)
  • Configure Galaxy for using these services
  • Install tools and workflows using the bioblend API.

The code of the GalaxyKickStart playbook is freely available at the ARTbio GitHub Repository https://github.com/ARTbio/GalaxyKickStart.

Deployment

  • start a GCE VM with the following characteristics

Google Instance for Ansible deployment

  • Name: ansible-galaxy
  • Region europe-west6 (Zurich) (or any region available with you Google coupon)
  • Zone: europe-west6-a (or -b or -c)
  • Configuration de la machine
    • OPTIMISEE POUR LE CALCUL
    • Série: C2
    • Type de machine: c2-standard-16 (16 processeurs virtuels, 64 Go de mémoire)
  • Disque de démarrage (Modifier)
    • IMAGES PUBLIQUES
    • Système d'exploitation: Ubuntu
    • Version*: Ubuntu 20.04 LTS
    • Type de disque de démarrage: Disque persistant avec équilibrage
    • Taille (Go): 200
    • SELECTONNER
  • Pare-feu
    • Check Autoriser le trafic HTTP
  • connect to you VM using the Google ssh console
  • start an interactive session as root using the command 
    sudo -i
  • the GalaxyKickstart ansible playbook is downloading automatically several Gb of cached resources in order to accelerate the deployment of the Galaxy server. Since you are going to do this ~18 times, it is better to distribute the cache downloads on several servers. Thus, we are going to divide the class in 3 groups, 🍩, 🍨 and 🍬, which will use each a deployment script obtaining the cached resources from 3 different servers located in Sorbonne-Université, AWS-Paris and Google-Europe, respectively.
  • Get the deployment script on your VM using the command:
for the team 🍩

Only for the team 🍩 

wget https://raw.githubusercontent.com/ARTbio/Run-Galaxy/master/deployment_scripts/run_ansible_analyse_genomes_2021-F.sh

for the team 🍨

Only for the team 🍨 

wget https://raw.githubusercontent.com/ARTbio/Run-Galaxy/master/deployment_scripts/run_ansible_analyse_genomes_2021-M.sh

for the team 🍬

Only for the team 🍬 

wget https://raw.githubusercontent.com/ARTbio/Run-Galaxy/master/deployment_scripts/run_ansible_analyse_genomes_2021-IEL.sh

  • We are now ready to run these scripts using the same command. In addition, all trainees may participate to the Pasteur 2021 Ansible Racing. In order to participate, you'll just have to put the time command just before the script invokation, as follows:
time sh run_ansible_analyse_genomes_2021*.sh

The Ultimate Pasteur 2021 Ansible Racing

Please copy the time info returned by your console at the end of the deploymment. It shoud look like this: 

real    37m23.924s
user    17m26.569s
sys     2m33.091s
Then Paste this time as a comment in this GitHub issue

  • When the deployment is finished, connect to your ansible-deployed "GalaxyKickStart" instance:

    Just click on the url displayed in your Google Cloud Engine Console.

  • Connect to your server as an admin:

    This time, ansible and the GalaxyKickStart playbook already programmatically registered an admin user. Just use the admin@galaxy.org:artbio2020 as credentials (user:password)

    When logged in, see that required tools as well as workflows are already installed !

Warning

artbio2020 is not really a decent password, please c h a n g e . y o u r . p a s s w o r d to avoid your Galaxy server getting hacked before the end of the course 😉