The document discusses Jenkins 2.0 and outlines goals to position Jenkins as a continuous delivery platform. It proposes three pillars for Jenkins 2.0: updated messaging and website to focus on continuous delivery, an improved out-of-the-box experience centered around workflow and pipeline functionality, and targeted internal changes like library upgrades and deprecated API removal. Potential pitfalls are also identified around scope, disruption, backwards compatibility and massive internal changes.
assembly a complex CD workflow today is like a LEGO model assembly
Workflow does this in one job
and in text / not just calling other jobs, but do stuff within itself
Tolerate loss of master & slaves
and in text / not just calling other jobs, but do stuff within itself
and in text / not just calling other jobs, but do stuff within itself
As an architect of the CD pipeline, you can define them as code in one place. Then product teams get to reuse them
Docker is driving changes in software development, and Jenkins is no exception. I want to discuss the continuous delivery in the world of containers
Let’s start with the simplest use case, containerizing app / how hard can it be? / Even simple things like this can be made simpler / Credentials management
Common credential across all the docker plugins
Your container building often depends on another container.
trigger plugin is also useful / ex. JIRA + LDAP
If a test pass, what do you do? You want to publish images to another repository / tag.
You’ve got stuff coming in, test it, reject it, test it, accept it / the same problem with these guys /
Not a new problem, the same with files / but Docker is worse
We need our own tracking tags
Tags get scanned everywhere / find where it is / And what we released so far is just a starter. We are going to expand it
for example, know what went into that container / important view point because it lets you focus on how things are flowing through the system
I did first handful of plugins myself
We know how to assemble LEGO / our pain points are often different.
Need to have enough to justify 2.0 but not everything for everyone / direction setting
We know how to assemble LEGO / our pain points are often different.
We know how to assemble LEGO / our pain points are often different.
We know how to assemble LEGO / our pain points are often different.
We know how to assemble LEGO / our pain points are often different.
We know how to assemble LEGO / our pain points are often different.
References to images get baked somewhere / these ideally live in a VCS / trigger activities in Jenkins
And then you start putting things together / the whole thing starts to look eerily similar.
Artifacts are different, languages are different, but they all follow the same paradigm.