zyj0808 / ansible-role-mutalyzer

Ansible role for Mutalyzer

Ansible role for Mutalyzer

This role deploys the HGVS variant nomenclature checker Mutalyzer.

New here? Check out How to use the Mutalyzer Ansible role.

Table of contents

• Requirements
• Description
• Dependencies
• Variables
• Local deployment with Vagrant

Requirements

• Debian 8 (Jessie) with a configured MTA
• Ansible version 2.0.1

Description

The deployment uses the following tools:

• nginx as reverse proxy for Gunicorn and for serving static files.
• systemd for process control.
• Gunicorn as WSGI HTTP server.
• virtualenv for isolation of the Mutalyzer package and its Python dependencies.
• PostgreSQL for the database (can be customized).
• Redis for stat counters and other in-memory stores (can be customized).

Three applications are served by nginx:

• Website: https://<servername>/
• HTTP/RPC+JSON webservice: https://<servername>/json
• SOAP webservice: https://<servername>/services

Files and directories

Logging is done at several levels in this stack:

• Mutalyzer log: /opt/mutalyzer/log/mutalyzer.log
• Mutalyzer website Gunicorn log: /opt/mutalyzer/versions/*/log/website.log
• Mutalyzer HTTP/RPC+JSON webservice Gunicorn log: /opt/mutalyzer/versions/*/log/service-json.log
• Mutalyzer SOAP webservice Gunicorn log: /opt/mutalyzer/versions/*/log/service-soap.log
• nginx access and error logs: /var/log/nginx/
• redis server logs: /var/log/redis/
• PostgreSQL server logs: /var/log/postgresql/

Tool configurations can be found here (but you should never manually touch them):

• Mutalyzer configuration: /opt/mutalyzer/versions/*/conf/settings.py
• Mutalyzer website Gunicorn configuration: /opt/mutalyzer/versions/*/conf/website.conf
• Mutalyzer HTTP/RPC+JSON webservice Gunicorn configuration: /opt/mutalyzer/versions/*/conf/service-json.conf
• Mutalyzer SOAP webservice Gunicorn configuration: /opt/mutalyzer/versions/*/conf/service-soap.conf
• nginx configuration: /etc/nginx/sites-available/mutalyzer-*

All Mutalyzer processes run as user mutalyzer, which is also the owner of everything under /opt/mutalyzer. Some other Mutalyzer related locations are:

• Mutalyzer cache: /opt/mutalyzer/cache/
• Mutalyzer Git clone: /opt/mutalyzer/src/mutalyzer/
• Mutalyzer versions: /opt/mutalyzer/versions/

Zero-downtime deployments

In order to obtain zero-downtime deployments, we fix and isolate deployment versions. A version is identified by its Git commit hash and contains a Python virtual environment, configuration files, log files, and unix sockets for the website and webservices. Several versions can co-exist, but only one version is active (published).

The deployment of a new version is done when the Mutalyzer Git repository checkout changes:

1. Create the new version in /opt/mutalyzer/versions/<git commit hash>/.
2. Run unit tests (if enabled, see variables below).
3. Run database migrations.
4. Start the new Gunicorn website and services.
5. Test the availability of the website and services.
6. Stop running batch processor and start a new one.
7. Reload nginx with the new Gunicorn upstreams.
8. Stop the old Gunicorn website and services (first completing all their pending requests).

In step 7, nginx will complete all existing requests from the old configuration while accepting requests with the new configuration, so this is zero-downtime.

The order above also means database migrations must always keep compatibility with the existing codebase, so some may have to be broken down into several steps and completed over several deployments.

Migrations that should be completed over several deployments are grouped per Mutalyzer release. It is therefore advised to always update Mutalyzer one release at a time and never skip a release to ensure database consistency and zero-downtime deployments. This can be controled by setting the mutalyzer_git_branch role variable to a release tag.

User environment

The ~/.bashrc file for user mutalyzer activates the Mutalyzer Python virtual environment and sets the MUTALYZER_SETTINGS environment variable. Administrative work is best done as that user:

sudo -u mutalyzer -i

Please note that you should re-source ~/.bashrc in any existing shell sessions after deploying a new Mutalyzer version, to switch to the current virtual environment.

Dependencies

This role depends on the following roles:

postgresql (optional)

https://git.lumc.nl/humgen-devops/ansible-role-postgresql

Variable overrides:

postgresql_databases:
  - name: mutalyzer
    encoding: UTF8
    lc_collate: 'en_US.UTF-8'
    lc_ctype: 'en_US.UTF-8'
    backup: true
postgresql_users:
  - name: mutalyzer
    password: "{{ mutalyzer_database_password }}"
    attributes: NOSUPERUSER,NOCREATEDB
    databases:
      - name: mutalyzer
        privileges: ALL

This role is only needed when the mutalyzer_database_url variable is null (default).

redis

https://git.lumc.nl/humgen-devops/ansible-role-redis

This role is only needed when the mutalyzer_redis_url variable is null (default).

nginx

https://git.lumc.nl/humgen-devops/ansible-role-nginx

mail-service-status

https://git.lumc.nl/humgen-devops/ansible-role-mail-service-status

Variables

Also see variables of dependencies.

mutalyzer_certificate

Default: localhost-insecure.crt (self-signed certificate for localhost)

SSL certificate file.

mutalyzer_certificate_key

Default: localhost-insecure.key

SSL certificate keyfile.

mutalyzer_database_url

Default: null

URL to use for connecting to a database in a form accepted by SQLAlchemy (see SQLAlchemy Database Urls).

If null, a local PostgreSQL database is used (managed by this role). If a database server is specified, it should be running when applying this role.

mutalyzer_database_password

Default: insecure_password

Password for the database user. Only used when mutalyzer_database_url is null (the password is part of the URL otherwise).

mutalyzer_redis_url

Default: null

URL to use for connecting to a Redis server in a form accepted by redis-py (see redis-py documentation).

If null, a local Redis server is used (managed by this role). If another Redis server is specified, it should be running when applying this role.

mutalyzer_server_name

Default: localhost

Server name by which Mutalyzer can be reached.

mutalyzer_website_proxy_read_timeout

Default: 60

Nginx read timeout for the website Gunicorn upstream.

mutalyzer_website_worker_class

Default: sync

Type of Gunicorn worker for the website. Must be one of sync, eventlet, gevent.

mutalyzer_website_workers

Default: 2

Number of Gunicorn workers for the website.

mutalyzer_website_timeout

Default: 30

Timeout before killing silent Gunicorn workers for the website.

mutalyzer_service_json_proxy_read_timeout

Default: 60

Nginx read timeout for the HTTP/RPC+JSON webservice Gunicorn upstream.

mutalyzer_service_json_worker_class

Default: sync

Type of Gunicorn worker for the HTTP/RPC+JSON webservice. Must be one of sync, eventlet, gevent.

mutalyzer_service_json_workers

Default: 1

Number of Gunicorn workers for the HTTP/RPC+JSON webservice.

mutalyzer_service_json_timeout

Default: 30

Timeout before killing silent Gunicorn workers for the HTTP/RPC+JSON webservice.

mutalyzer_service_soap_proxy_read_timeout

Default: 60

Nginx read timeout for the SOAP webservice Gunicorn upstream.

mutalyzer_service_soap_worker_class

Default: sync

Type of Gunicorn worker for the SOAP webservice. Must be one of sync, eventlet, gevent.

mutalyzer_service_soap_workers

Default: 1

Number of Gunicorn workers for the SOAP webservice.

mutalyzer_service_soap_timeout

Default: 30

Timeout before killing silent Gunicorn workers for the SOAP webservice.

mutalyzer_batch_notification_email

Default: noreply@localhost

Address used as sender in batch job notification emails.

mutalyzer_max_cache_size

Default: 52428800 (50 MB)

Maximum size of the cache directory (in bytes).

mutalyzer_extractor_max_input_length

Default: 50000 (50 Kbp)

Maximum sequence length for the description extractor (in bases).

mutalyzer_remote_caches

Default: []

List of remotes to sync the cache with, where each remote is a dictionary with the following fields:

• name: Remote name (should contain only letters, digits, -, and _).
• wsdl: Location of the remote WSDL description.
• url_template: URL for remote downloads, in which the filename is to be substituted for {file}.

mutalyzer_piwik

Default: null

Piwik web analytics configuration as a dictionary with the following fields:

• base_url: Piwik server base URL (include protocol, no trailing slash).
• site_id: Piwik site ID.

mutalyzer_prune_versions

Default: true

Whether or not to remove old Mutalyzer versions, including their Python virtual environment, log files, and configuration.

mutalyzer_unit_tests

Default: false

Whether or not to run Mutalyzer unit tests. Deployment will be aborted if they fail.

mutalyzer_git_repository

Default: https://github.com/mutalyzer/mutalyzer.git

Mutalyzer Git repository URL to clone.

mutalyzer_git_branch

Default: master

Mutalyzer Git repository branch to checkout.

Local deployment with Vagrant

Easy deployment on a local virtual machine using Vagrant is provided in the vagrant directory.