Skip to content

Latest commit

 

History

History
69 lines (40 loc) · 2.48 KB

CI-CD.md

File metadata and controls

69 lines (40 loc) · 2.48 KB
Raw

Introduction

This document describes Softcatalà's Gitlab CI/CD system to deploy the system in production.

Pipelines

github/translation-memory-tools

Responsibilities:

  • Builds the image that will fetch the translation memories, generates terminology and quality reports
  • Runs the image that fetches the translation memories

Storage:

  • The docker volume volume /srv/volumes/tmt-files is used to preserve the data between executions of the pipeline.

Once the run job finishes, a Docker image based on softcatala/sc-static-file-server is created to store all the data.

This image serves two purposes:

  • Contains the information that will be used by the web app
  • It contains the static files which will be served at https://static.softcatala.org. These static files include:
    • the HTML quality reports,
    • the terminology web/csv
    • files and the translation memories (PO/TMX).
    • the index_data.json to later create a searchable index
    • the glossary.json to later create a searchable glossary / terminology

translation-memory-tools-webservice

Responsibilities:

  • Create the Whoosh searchable index
  • Builds the web app
  • Executes integration tests

docker/translation-memory-tools

Contains the docker-compose.yml used to run the system in production.

Common tasks and troubleshooting with CI/CD

Deploy a Git branch

There is no staging environment, you need to deploy your branches in pro.

To do this:

Rolling back to a previous version

Go to https://gitlab.softcatala.org/docker/translation-memory-tools and change the docker-compose.yml file to a previous branch.

All the branches at tagged also with the date, such as 20231115-100154 to enable to rollback to a specific date.

Downloading the generated data image from translation-memory-tools

In the translation-memory-tools repo, use this Makefile task:

make docker-run-gitlab-data-debug

The image will be downloaded locally and you will login into the image which will allow you to inspect the data image generated