gitlab.rst

Using in Gitlab CI

Vagga's defaults are very conservative and are better suited for workstations. For higher performance on CI you need some settings activated.

For example, here is good settings for Gitlab CI:

storage-dir: /var/lib/gitlab-runner/containers
storage-subdir-from-env-var: CI_PROJECT_NAME
cache-dir: /var/lib/gitlab-runner/cache
ubuntu-mirror: http://ua.archive.ubuntu.com/
alpine-mirror: http://dl-cdn.alpinelinux.org/alpine/
build-lock-wait: true
index-all-images: true
hard-link-identical-files: true
ubuntu-skip-locking: true
hard-link-between-projects: true
versioned-build-dir: true

Note: some of these settings are non-secure if single gitlab runner is used across organizations. See :ref:`Settings` for more info on each setting.

Make sure to create dirs:

mkdir /var/lib/gitlab-runner/containers
mkdir /var/lib/gitlab-runner/cache

Also check that subuid/subgid were allocated, the following line or equivalent must be both in /etc/subuid and /etc/subgid:

gitlab-runner:200000:65536

Gitlab-runner config:

concurrent = 3
check_interval = 0

[[runners]]
  name = "vagga runner"
  url = "https://lab.thinkglobal.space"
  token = "xxx_yourToken_xxx"
  executor = "shell"
  pre_build_script = "vagga --version"
  [runners.cache]

Note: concurrency > 1 is okay, pre_build_script shows vagga version

Also it's a good idea to make cleanup regularly:

vagga _clean --global --unused --al-least 7days

(run as gitlab-runner user)