Open main menu

Automated deployment of MediaWiki

Revision as of 11:14, 24 August 2020 by Admin (talk | contribs) (reorganize)

As a contributor to Meza, and developer of QualityBox, we've been closely following the efforts in the MediaWiki community around easier/automated deployments of MediaWiki. Actually, it's a goal of the MWStakeholders Group. And, it's a crucial aspect of the whole 2017 Wikimedia Technology Department Goals since at least 2018. Here is a list of some of the efforts and projects within and outside the WMF for automating deployments of MediaWiki:

WMF

For Distributors

MediaWiki with Kubernetes

MediaWiki on Docker


Node Services

  • MediaWiki Node Services Run several MediaWiki nodejs services (RESTBase, Parsoid) in a single docker container. Useful for small / low-memory installs. By WMF's d00rman (Marko Obrovac - Croatia) and gwicke (Gabriel Wicke - Menlo Park)

Lex Sulzer


Configuration Management and Deployment at WMF

It should be noted that the WMF uses Puppet for configuration management (See the operations/mediawiki-config repo), and Scap3 (which is used in SWAT) (obsolete) for deployments. There is the Cloud VPS which is powered by wp:OpenStack for projects related to the Wikimedia movement. The environment includes access to a variety of data services. Cloud VPS is meant to make it easier for developers and system administrators to try out improvements to Wikimedia infrastructure (including MediaWiki), power research and analytics, and host projects that are not viable in the Toolforge environment. Virtual machines running on Cloud VPS are subject to the instance lifecycle.

Puppet


Config Settings

With any automation tooling and deployment, you will inevitably need to refer to the long list of MediaWiki Configuration settings.