I have a dockerized app and I set up a CI pipeline for it using CircleCI.
In this workflow, I have tests that if passed, will trigger the build of the docker image and push it to DockerHub.
Then, I have this Linode server where I manually pull this docker image from DockerHub.
The problem is every time I commit code, I also need to wait for the CI workflows to finish so that I can manually update the docker image in Linode.
Is there any way I can automate this so that I can add Continuous Delivery from CircleCI to Linode?
Related
Currently I trigger a Cloud Build each time a pull request is completed.
The image is built correctly, but we have to manually go to Edit and Deploy New Revision and select the most recent docker image to deploy.
How can we automate this process and have the a container deployed from the image automatically?
You can do it with a pretty simple GitHub Action. I have followed this article:
https://towardsdatascience.com/deploy-to-google-cloud-run-using-github-actions-590ecf957af0
Cloud Run also natively integrates with Cloud Build. You can import a GitHub repository and it sets up a GCB trigger for your repository (on the specified branch or tag filter).
You can use GitLab CI to automate your Cloud Run deployment.
Here are the tutorial if you want to automate your deployment with GitLab CI Link
We are trying to setup a development CI/CD pipeline with Jenkins that builds the Docker Images and deploy that Directly to AWS EKS cluster. Is this even possible??
Our Existing system
Jenkins as CI to pick the Code from GitLab and Build Docker Image
After Build, Jenkins push the Image to Jfrog Artifactory(Professional)
We use Harness for CD, that picks the Image from Artifactory and deploy that
to AWS
Here, Artifiactory and Harness Incurs cost for us and we don't want to use that for Development builds. So, we have setup a Docker Registry with Soantype Nexus3 OSS(open source version).
I would like to know two Options here:
if I can use Jenkins to Build Docker Image and Push that to Nexus Docker Registry and Use Jenkins Itself for CD to deploy that to AWS EKS?
Build Docker images with Jenkins and directly deploy that to AWS EKS without even having to store it in a docker registry?
Any suggestions and help is highly appreciated!
the first option much better.
because one day may need roll-back docker image on Kubernetes. (even development environment)
or you can use AWS ECR. it's easier to use on EKS.
and I think ECR is cheaper than Nexus operation cost.
You may be happy to know that Harness has created a free software version of it's CD service, called Harness Continuous Delivery Community Edition, which should work nicely for your development builds.
Having built,run and executed tests against a docker image on a CI build server(TeamCity2017), how should we deploy it to further machines?
How, for example, if we push it to a Docker registry, would our CI server instruct the target machine to pull and run the image? I.e. where it an application we would use Octopus for this deployment step, but our Octopus server doesn't support Docker deployments as yet.
Any guidance appreciated.
Michael McD.
I would use Octo to deploy your images onto target machines. You'd need to use powershell scripts to have your machines run the images. Or you can use something like Rancher, which is a docker swarm manager. There is no feasible way to have TeamCity deploy your images. The software simply isn't built to be able to do deploys.
The Rancher solution would not be automated, at least not to my knowledge. You would have to trigger upgrades when a new image is pushed to the docker registry.
I have a github-repository, that is linked to automated build on Docker. Consequently, on each commit to master-branch, docker triggers building of Docker-image.
Also, each commit is tested by Travis CI automatically.
My question is: is there any way to trigger Docker only if travis finishes successfully? Do I need some sort of webhook or something like that for my goal?
You could trigger the Travis CI test after the repository is pushed. Then, in the deploy step you could trigger a build on Docker. Or even do the build inside Travis, and just push the image to the repository you are using.
Travis has a nice overview of how to make this flow happen here.
The gist is that you're going to need to have sudo: required, so you're going to be running in a VM instead of inside Docker, as is the standard way in Travis. You also need to add docker as a service, much like you'd add redis or postgres for an integration test. The Pushing Docker Image to a Registry section has a lot of info on setting things up for the actual deployment. I'd use an actual deploy step with the script provider, rather than after_success, but that's up to you.
I've been trying gitlab and its CI workflow these days, but found myself confused when I saw these messages during a build:
gitlab-ci-multi-runner 0.6.2 (3227f0a)
Using Docker executor with image mydocker:latest ...
Running on runner-5498280b-project-20053-concurrent-0 via jls-MacBook-Pro...
I register a project specified runner instead of using the shared ones.
Is gitlab actually running all the CI build process via my own machine? What if other co-workers push to this project while my computer was off? I just thought that gitlab would provide every project with a cloud CI server... So , I don't want to turn my own computer into such a server. Am I missing something on its docs?
See this Gitlab Docs
Is gitlab actually running all the CI build process via my own machine?
--Yes, Because you specified it. If you don't want to run into your machine, Then deploy the docker to other server.
What if other co-workers push to this project while my computer was off?
--Then, no build will going to happen,