#dind — Public Fediverse posts

GitHub self-hosted runners в Docker: как поднять несколько изолированных раннеров на одном хосте. Часть 1

Если у вас приватные репозитории на GitHub и команда, которая регулярно упирается в лимит времени GitHub Actions, эта статья сэкономит вам пару недель экспериментов. Рассказываем, как мы подняли self-hosted раннеры в Docker, настроили их репликацию через Docker Compose и почему в итоге пришли к Docker-in-Docker. Разобрали подробно, по шагам: эволюционировали от bare metal раннеров до докеризованной конфигурации. Репозиторий с сетапом находится в открытом доступе.

https://habr.com/ru/articles/1047640/

#docker #dockercompose #github_actions #cicd #github #dind #runner #runners

How is this whole #ansible #molecule tool so bad.

It has mountains of documentation examples that literally do not work at all.

Mentions a whole slew of configuration parameters that seem entirely ignored.

And worst of all, it forces a very particular directory structure that makes sharing these huge piles of boilerplate across roles and playbooks, impossible...

But sadly it's the only tool in human existence to seamlessly allow dind testing with systemd based containers.

#dind #systemd

Кэш-монтирование или Cache mount в Docker. Использование кэш-монтирования в GitLab CI/CD

Кэш-монтирование в Docker — это мощный инструмент, который может значительно ускорить процесс сборки образов в CI/CD. Но его правильная настройка в GitLab требует понимания не только работы BuildKit, но и особенностей взаимодействия с различными конфигурациями GitLab Runner — DinD и DooD. В этой статье я расскажу об устройстве кэш-монтирования в Docker и как его использовать в GitLab CI/CD. На примере простой сборки Python-проекта покажу готовые решения для кэширования зависимостей, чтобы оптимизировать сборку Docker-образов.

https://habr.com/ru/articles/995528/

#docker #deploy #gitlab #python #cache #dockerfile #cicd #dind #buildkit #git

This is so damn cool. Relatively secure docker in docker with very an isolated docker daemon context and low overhead. A great alternative to MicroVM's like #Firecracker.

https://www.docker.com/resources/docker-in-docker-containerized-ci-workflows-dockercon-2023/

I love the presentation style too. I followed along in my terminal and everything worked. So much more helpful for me than "go to this website to learn more".

#dind #microvm

@deliverance If #Kaniko is not your speed, You can still do better than #DinD by using #PinD or even #PinP

How to use Podman inside of a container
https://www.redhat.com/sysadmin/podman-inside-container

This way you don't need a daemon, and you can run rootless.

@deliverance If #Kaniko is not your speed, You can still do better than #DinD by using #PinD or even #PinP

How to use Podman inside of a container
https://www.redhat.com/sysadmin/podman-inside-container

This way you don't need a daemon, and you can run rootless.

@fyw321 @geerlingguy My 8 node cluster costs around 55W via POE. It is 3 pi4 8G control+worker nodes, and 5 pi4 4G worker-only nodes (1 is actually 8G.)

Storage is #iscsi on spinning rust on #TrueNAS, but you can do all this on local disks.

It handles a LOT:
- #Promstack, #FluxCD, #Calico, various scrapers
- #Adguard
- #Blocky (another adblocking #dns)
- #Jellyfin
- #Ombi, Prowlarr, *arr, Deluge
- Home automation helpers (#ser2sock instances, #zigbee/#zwave 2mqtt, but not HA itself)
- #Argo for builds (deprecated in favor of the x64 cloud lab. Building x64 docker containers on arm is BAD)
- Democratic CSI for iscsi/nfs
- #GoHarbor container registry & #dockerhub cache
- #Mealie recipe manager
- #Monica contact manager
- #SMTP relay to gmail
- #Ubiquiti console (bootstrapping becomes a chicken and egg problem though, if it goes down wrong)
- #Wireguard #VPN server
- #Whoogle
- #VisualStudioCode and a #dind sidecar
- SSL termination for most of the rest of the network

@fyw321 @geerlingguy My 8 node cluster costs around 55W via POE. It is 3 pi4 8G control+worker nodes, and 5 pi4 4G worker-only nodes (1 is actually 8G.)

Storage is #iscsi on spinning rust on #TrueNAS, but you can do all this on local disks.

It handles a LOT:
- #Promstack, #FluxCD, #Calico, various scrapers
- #Adguard
- #Blocky (another adblocking #dns)
- #Jellyfin
- #Ombi, Prowlarr, *arr, Deluge
- Home automation helpers (#ser2sock instances, #zigbee/#zwave 2mqtt, but not HA itself)
- #Argo for builds (deprecated in favor of the x64 cloud lab. Building x64 docker containers on arm is BAD)
- Democratic CSI for iscsi/nfs
- #GoHarbor container registry & #dockerhub cache
- #Mealie recipe manager
- #Monica contact manager
- #SMTP relay to gmail
- #Ubiquiti console (bootstrapping becomes a chicken and egg problem though, if it goes down wrong)
- #Wireguard #VPN server
- #Whoogle
- #VisualStudioCode and a #dind sidecar
- SSL termination for most of the rest of the network

It's just that the test setup is cursed:

The tests run on #GitLab runners using the #Docker executor in privileged mode. This enables us to start a #KVM based #AVD for Android tests and to deploy all individual homeserver implementations using #DinD.

In order to get better debugging information, the headless virtual screen is recorded via #VNC in order to record all ongoing tests for later use.

Some people told us they don't like our tests.