Move developer documentation to docs/devel/

Fix links.
duda · Oct 15, 2014 · 2ce16ac · 2ce16ac
1 parent 3f52a97
commit 2ce16ac
Show file tree

Hide file tree

Showing 9 changed files with 648 additions and 2 deletions.
diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md
@@ -21,7 +21,7 @@ Follow either of the two links above to access the appropriate CLA and instructi
 
 ## Protocols for Collaborative Development
 
-Please read [this doc](docs/collab.md) for information on how we're running development for the project.
+Please read [this doc](docs/devel/collab.md) for information on how we're running development for the project.
 
 ## Adding dependencies
 

diff --git a/README.md b/README.md
@@ -46,7 +46,7 @@ Check out examples of Kubernetes in action, and community projects in the larger
 * [Example of dynamic updates](examples/update-demo/README.md)
 * [Cluster monitoring with heapster and cAdvisor](https://github.com/GoogleCloudPlatform/heapster)
 * [Community projects](https://github.com/GoogleCloudPlatform/kubernetes/wiki/Kubernetes-Community)
-* [Development guide](docs/development.md)
+* [Development guide](docs/devel/development.md)
 
 Or fork and start hacking!
 

diff --git a/docs/devel/collab.md b/docs/devel/collab.md
@@ -0,0 +1,37 @@
+# On Collaborative Development
+
+Kubernetes is open source, but many of the people working on it do so as their day job.  In order to avoid forcing people to be "at work" effectively 24/7, we want to establish some semi-formal protocols around development.  Hopefully these rules make things go more smoothly.  If you find that this is not the case, please complain loudly.
+
+## Patches welcome
+
+First and foremost: as a potential contributor, your changes and ideas are welcome at any hour of the day or night, weekdays, weekends, and holidays.  Please do not ever hesitate to ask a question or send a PR.
+
+## Timezones and calendars
+
+For the time being, most of the people working on this project are in the US and on Pacific time.  Any times mentioned henceforth will refer to this timezone.  Any references to "work days" will refer to the US calendar.
+
+## Code reviews
+
+All changes must be code reviewed.  For non-maintainers this is obvious, since you can't commit anyway.  But even for maintainers, we want all changes to get at least one review, preferably from someone who knows the areas the change touches.  For non-trivial changes we may want two reviewers.  The primary reviewer will make this decision and nominate a second reviewer, if needed.  Except for trivial changes, PRs should sit for at least a 2 hours to allow for wider review.
+
+Most PRs will find reviewers organically.  If a maintainer intends to be the primary reviewer of a PR they should set themselves as the assignee on GitHub and say so in a reply to the PR.  Only the primary reviewer of a change should actually do the merge, except in rare cases (e.g. they are unavailable in a reasonable timeframe).
+
+If a PR has gone 2 work days without an owner emerging, please poke the PR thread and ask for a reviewer to be assigned.
+
+Except for rare cases, such as trivial changes (e.g. typos, comments) or emergencies (e.g. broken builds), maintainers should not merge their own changes.
+
+Expect reviewers to request that you avoid [common go style mistakes](https://code.google.com/p/go-wiki/wiki/CodeReviewComments) in your PRs.
+
+## Assigned reviews
+
+Maintainers can assign reviews to other maintainers, when appropriate. The assignee becomes the shepherd for that PR and is responsible for merging the PR once they are satisfied with it or else closing it.  The assignee might request reviews from non-maintainers.
+
+## Merge hours
+
+Maintainers will do merges between the hours of 7:00 am Monday and 7:00 pm (19:00h) Friday.  PRs that arrive over the weekend or on holidays will only be merged if there is a very good reason for it and if the code review requirements have been met.
+
+There may be discussion an even approvals granted outside of the above hours, but merges will generally be deferred.
+
+## Holds
+
+Any maintainer or core contributor who wants to review a PR but does not have time immediately may put a hold on a PR simply by saying so on the PR discussion and offering an ETA measured in single-digit days at most.  Any PR that has a hold shall not be merged until the person who requested the hold acks the review, withdraws their hold, or is overruled by a preponderance of maintainers.
diff --git a/docs/devel/development.md b/docs/devel/development.md
@@ -0,0 +1,179 @@
+# Development Guide
+
+# Releases and Official Builds
+
+Official releases are built in Docker containers.  Details are [here](build/README.md).  You can do simple builds and development with just a local Docker installation.  If want to build go locally outside of docker, please continue below.
+
+## Go development environment
+
+Kubernetes is written in [Go](http://golang.org) programming language. If you haven't set up Go development environment, please follow [this instruction](http://golang.org/doc/code.html) to install go tool and set up GOPATH. Ensure your version of Go is at least 1.3.
+
+## Put kubernetes into GOPATH
+
+We highly recommend to put kubernetes' code into your GOPATH. For example, the following commands will download kubernetes' code under the current user's GOPATH (Assuming there's only one directory in GOPATH.):
+
+```
+$ echo $GOPATH
+/home/user/goproj
+$ mkdir -p $GOPATH/src/github.com/GoogleCloudPlatform/
+$ cd $GOPATH/src/github.com/GoogleCloudPlatform/
+$ git clone git@github.com:GoogleCloudPlatform/kubernetes.git
+```
+
+The commands above will not work if there are more than one directory in ``$GOPATH``.
+
+(Obviously, clone your own fork of Kubernetes if you plan to do development.)
+
+## godep and dependency management
+
+Kubernetes uses [godep](https://github.com/tools/godep) to manage dependencies. It is not strictly required for building Kubernetes but it is required when managing dependencies under the Godeps/ tree, and is required by a number of the build and test scripts. Please make sure that ``godep`` is installed and in your ``$PATH``.
+
+### Installing godep
+There are many ways to build and host go binaries. Here is an easy way to get utilities like ```godep``` installed:
+
+1. Ensure that [mercurial](http://mercurial.selenic.com/wiki/Download) is installed on your system. (some of godep's dependencies use the mercurial
+source control system).  Use ```apt-get install mercurial``` or ```yum install mercurial``` on Linux, or [brew.sh](http://brew.sh) on OS X, or download
+directly from mercurial.
+2. Create a new GOPATH for your tools and install godep:
+```
+export GOPATH=$HOME/go-tools
+mkdir -p $GOPATH
+go get github.com/tools/godep
+```
+
+3. Add $GOPATH/bin to your path. Typically you'd add this to your ~/.profile:
+```
+export GOPATH=$HOME/go-tools
+export PATH=$PATH:$GOPATH/bin
+```
+
+### Using godep
+Here is a quick summary of `godep`.  `godep` helps manage third party dependencies by copying known versions into Godeps/_workspace.  You can use `godep` in three ways:
+
+1. Use `godep` to call your `go` commands.  For example: `godep go test ./...`
+2. Use `godep` to modify your `$GOPATH` so that other tools know where to find the dependencies.  Specifically: `export GOPATH=$GOPATH:$(godep path)`
+3. Use `godep` to copy the saved versions of packages into your `$GOPATH`.  This is done with `godep restore`.
+
+We recommend using options #1 or #2.
+
+## Hooks
+
+Before committing any changes, please link/copy these hooks into your .git
+directory. This will keep you from accidentally committing non-gofmt'd go code.
+
+```
+cd kubernetes
+ln -s hooks/prepare-commit-msg .git/hooks/prepare-commit-msg
+ln -s hooks/commit-msg .git/hooks/commit-msg
+```
+
+## Unit tests
+
+```
+cd kubernetes
+hack/test-go.sh
+```
+
+Alternatively, you could also run:
+
+```
+cd kubernetes
+godep go test ./...
+```
+
+If you only want to run unit tests in one package, you could run ``godep go test`` under the package directory. For example, the following commands will run all unit tests in package kubelet:
+
+```
+$ cd kubernetes # step into kubernetes' directory.
+$ cd pkg/kubelet
+$ godep go test
+# some output from unit tests
+PASS
+ok      github.com/GoogleCloudPlatform/kubernetes/pkg/kubelet   0.317s
+```
+
+## Coverage
+```
+cd kubernetes
+godep go tool cover -html=target/c.out
+```
+
+## Integration tests
+
+You need an etcd somewhere in your PATH. To install etcd, run:
+
+```
+cd kubernetes
+hack/install-etcd.sh
+sudo ln -s $(pwd)/third_party/etcd/bin/etcd /usr/bin/etcd
+```
+
+```
+cd kubernetes
+hack/test-integration.sh
+```
+
+## End-to-End tests
+
+You can run an end-to-end test which will bring up a master and two minions, perform some tests, and then tear everything down. Make sure you have followed the getting started steps for your chosen cloud platform (which might involve changing the `KUBERNETES_PROVIDER` environment variable to something other than "gce".
+```
+cd kubernetes
+hack/e2e-test.sh
+```
+
+Pressing control-C should result in an orderly shutdown but if something goes wrong and you still have some VMs running you can force a cleanup with the magical incantation:
+```
+hack/e2e-test.sh 1 1 1
+```
+
+## Testing out flaky tests
+[Instructions here](docs/devel/flaky-tests.md)
+
+## Add/Update dependencies
+
+Kubernetes uses [godep](https://github.com/tools/godep) to manage dependencies. To add or update a package, please follow the instructions on [godep's document](https://github.com/tools/godep).
+
+To add a new package ``foo/bar``:
+
+- Make sure the kubernetes' root directory is in $GOPATH/github.com/GoogleCloudPlatform/kubernetes
+- Run ``godep restore`` to make sure you have all dependancies pulled.
+- Download foo/bar into the first directory in GOPATH: ``go get foo/bar``.
+- Change code in kubernetes to use ``foo/bar``.
+- Run ``godep save ./...`` under kubernetes' root directory.
+
+To update a package ``foo/bar``:
+
+- Make sure the kubernetes' root directory is in $GOPATH/github.com/GoogleCloudPlatform/kubernetes
+- Run ``godep restore`` to make sure you have all dependancies pulled.
+- Update the package with ``go get -u foo/bar``.
+- Change code in kubernetes accordingly if necessary.
+- Run ``godep update foo/bar`` under kubernetes' root directory.
+
+## Keeping your development fork in sync
+
+One time after cloning your forked repo:
+
+```
+git remote add upstream https://github.com/GoogleCloudPlatform/kubernetes.git
+```
+
+Then each time you want to sync to upstream:
+
+```
+git fetch upstream
+git rebase upstream/master
+```
+
+## Regenerating the API documentation
+
+```
+cd kubernetes/api
+sudo docker build -t kubernetes/raml2html .
+sudo docker run --name="docgen" kubernetes/raml2html
+sudo docker cp docgen:/data/kubernetes.html .
+```
+
+View the API documentation using htmlpreview (works on your fork, too):
+```
+http://htmlpreview.github.io/?https://github.com/GoogleCloudPlatform/kubernetes/blob/master/api/kubernetes.html
+```
diff --git a/docs/devel/flaky-tests.md b/docs/devel/flaky-tests.md
@@ -0,0 +1,52 @@
+# Hunting flaky tests in Kubernetes
+Sometimes unit tests are flaky.  This means that due to (usually) race conditions, they will occasionally fail, even though most of the time they pass.
+
+We have a goal of 99.9% flake free tests.  This means that there is only one flake in one thousand runs of a test.
+
+Running a test 1000 times on your own machine can be tedious and time consuming.  Fortunately, there is a better way to achieve this using Kubernetes.
+
+_Note: these instructions are mildly hacky for now, as we get run once semantics and logging they will get better_
+
+There is a testing image ```brendanburns/flake``` up on the docker hub.  We will use this image to test our fix.
+
+Create a replication controller with the following config:
+```yaml
+id: flakeController
+desiredState:
+  replicas: 24
+  replicaSelector:
+    name: flake
+  podTemplate:
+    desiredState:
+      manifest:
+        version: v1beta1
+        id: ""
+        volumes: []
+        containers:
+        - name: flake
+          image: brendanburns/flake
+          env:
+          - name: TEST_PACKAGE
+            value: pkg/tools
+          - name: REPO_SPEC
+            value: https://github.com/GoogleCloudPlatform/kubernetes
+      restartpolicy: {}
+    labels:
+      name: flake
+labels:
+  name: flake
+```
+
+```./cluster/kubecfg.sh -c controller.yaml create replicaControllers```
+
+This will spin up 100 instances of the test.  They will run to completion, then exit, the kubelet will restart them, eventually you will have sufficient
+runs for your purposes, and you can stop the replication controller:
+
+```sh
+./cluster/kubecfg.sh stop flakeController
+./cluster/kubecfg.sh rm flakeController
+```
+
+Now examine the machines with ```docker ps -a``` and look for tasks that exited with non-zero exit codes (ignore those that exited -1, since that's what happens when you stop the replica controller)
+
+Happy flake hunting!