Gitea and

Recently I’ve been frustrated at work while using Gitlab and Jenkins for various reasons, some of the integrations are really fragile due to some plugins we use, both Jenkins and Gitlab are incredibly bloated and use insane amounts of resources and they are simply not reasonable choices for a private setup. Also I recently replaced my server at home (basically a machine that does almost everything I want at home) from a 32 bit atom to an I5 intel machine with 16 gigs of RAM etc, which means I have totally different resources to work with. For example, 32 bit i386 cpu’s are not supported out of the box by docker, and the cpu was quite overloaded. With the new box I’ve been able to play around a little with my setup.

I’ve previously used just a basic and manual git setup at home, with approximately 110’ish repositories in it. I’ve been playing around with gitea and at home with the new server and am very pleased with it, even though it took a bit of work to get used to it. In all honesty, I’ve only done fairly basic work with it so far. The only really complex stuff I’ve done was to move my existing git repositories into the gitea environment by scripting a bit and using the gitea web based API.

Regarding gitea, so far I’ve noticed the following things:

  • Very slim by comparison to other options, currently uses around 60 MB of RAM and not a lot of CPU from what I’ve seen. Especially considering what you get.
  • UI is fairly similar to gitlab/github.
  • Setup was very simple except for database connection (I winded up just using sqlite3 I believe, I was lazy and also don’t expect more than a very few users).
  • Pull Request is really nice.
  • Issue tracking seems to work fairly well.
  • Docker setup with volumes is very easy.
  • Seems to have the essentials in plugins etc that I need.
  • API seems very nice, I’ve only used it for the migration so far though.
  • The only bad part I’ve seen so far is that the administration panel might be a bit spartan at times, but I don’t really mind.

Regarding, my first impressions are:

  • I absolutely love the yaml file format so far.
  • UI is incredibly clean, on the verge of too clean.
  • Integration with gitea was super simple once I actually got it working (documentation was not 100% accurate I think).
  • Simple to get started with if you have a sane build pattern.
  • Nice integration to gitea and you get marks on build statuses etc. Interesting to find if you can block a build from being merged based on build results as well.
  • I’ve managed to make a simple build of a playground project I have by adding a Dockerfile which is built into an image when the build is started, then a continuation of the build builds my project inside the docker image we just built.
  • First time using docker-compose so it was a bit of a hassle understanding this, but it was fun ;). Not always obvious where some configuration should be placed etc.
  • Yaml file format definitely not enough for the type of pipelines we do professionally though :(.
  • Pleasantly surprised you can actually add and remove build slaves to the platform.
  • Also pleasantly surprised by how to do parallel build steps. Syntax is super simple.
  • I really lack some form of artifact storage, or at least a plugin for something that is not either cloud based or incredibly enterprisey (artifactory). Actually, I’ve had issues just finding a good light weight open source artifact storage so far…
  • I also lack some form of nice presentation of various build artifacts, code coverage or unit test results etc.

In all, pleasantly surprised by how simple this was to setup and configure. It was a fun trip and I’ll continue using it at home for now.

As a sidenote, for the stuff I have on github, I do like to use travis, it also has a nice syntax and is a nice solution.

DevOps World Nice 2018

Long time again, I seem to be bad at updating this site at times.

I’ve recently moved into DevOps as a coincidence. I feel like this is what I’ve done for ages, but I just never set a name on it really. Also, this time it’s proper DevOps I’d say, a team of people maintaining build environments, CI/CD environments, gitlab, and so forth. It’s been a long time since I did proper server maintenance however, and it’s kind of interesting to see what happened. New tools, new possibilities, containerization, monitoring, virtualization, cloud, and so forth.

Because of said circumstances, I went to the DevOps World conference (previously Jenkins World I guess) in Nice and was pleasantly surprised by some of the development done to Jenkins and what they are trying to do. I’m a bit cautious with some of the changes though and what I expect of the future. One of my main issues from work is that a lot of the clients we have are very sensitive about where data is stored, to the point where they set up private clouds to keep stuff in house and in many cases blocking internet access almost completely. A lot of the changes in Jenkins and from CloudBees is about moving stuff into the public clouds from private installations. IE, something that we will not be able to do, even though my company is all for working as open as possible and driving our clients to open up what they are doing as well. There is obviously stuff you don’t want to open as a corporation, but a lot of it is not really specific to you and the help you can receive from collaborating with your competition is actually quite tremendous, hence driving down part costs etc.

So, all that said. I’m super happy to have been to the conference and have heard a lot new stuff. My main takeaways:

Jenkins x seems really interesting, it’s basically a stripped down Jenkins master running in kubernetes/cloud swarm but you spin up a new master for every build so the single point of failure is removed in it, and the kubernetes ecosystem for most of the extraneous systems such as web-hooks, etc. The bad part of this is that you loose the Jenkins UI  and Jenkins x is completely Command Line Interface based. CloudBees has some proprietary UI’s for Jenkins X however…

Configuration as Code Plugin seems really nice as I’d really like to easily take my Jenkins Master and create staging environments to test new changes easily. I will be experimenting with this most definitely.

Speeches about breaking down and splitting down your Jenkins Master and basically divide and conquer. Try to do different things on different servers basically, and don’t create a Jenkinstein with almost all the plugins in a single instance to cater to all types of jobs, etc.

Jenkins seems to be moving a lot into the public cloud, which unfortunately is bad for our industry as already mentioned. However, I’m really intrigued by the scaling possibilities that was shown using GCP and AWS for example.

Also, a lot of other good talks.

All this said, I’ve been experimenting with alternative CI and git hosting solutions, which I find rather interesting and I’ll write more about that in a close future…. 😉