b29d9cdbcf
Automatic merge from submit-queue
Build release tars using bazel
**What this PR does / why we need it**: builds equivalents of the various kubernetes release tarballs, solely using bazel.
For example, you can now do
```console
$ make bazel-release
$ hack/e2e.go -v -up -test -down
```
**Special notes for your reviewer**: this is currently dependent on 3b29803eb5, which I have yet to turn into a pull request, since I'm still trying to figure out if this is the best approach.
Basically, the issue comes up with the way we generate the various server docker image tarfiles and load them on nodes:
* we `md5sum` the binary being encapsulated (e.g. kube-proxy) and save that to `$binary.docker_tag` in the server tarball
* we then build the docker image and tag using that md5sum (e.g. `gcr.io/google_containers/kube-proxy:$MD5SUM`)
* we `docker save` this image, which embeds the full tag in the `$binary.tar` file.
* on cluster startup, we `docker load` these tarballs, which are loaded with the tag that we'd created at build time. the nodes then use the `$binary.docker_tag` file to find the right image.
With the current bazel `docker_build` rule, the tag isn't saved in the docker image tar, so the node is unable to find the image after `docker load`ing it.
My changes to the rule save the tag in the docker image tar, though I don't know if there are subtle issues with it. (Maybe we want to only tag when `--stamp` is given?)
Also, the docker images produced by bazel have the timestamp set to the unix epoch, which is not great for debugging. Might be another thing to change with a `--stamp`.
Long story short, we probably need to follow up with bazel folks on the best way to solve this problem.
**Release note**:
```release-note
NONE
```
Cluster add-ons
Cluster add-ons are resources like Services and Deployments (with pods) that are
shipped with the Kubernetes binaries and are considered an inherent part of the
Kubernetes clusters. The add-ons are visible through the API (they can be listed using
kubectl), but direct manipulation of these objects through Apiserver is discouraged
because the system will bring them back to the original state, in particular:
- If an add-on is deleted, it will be recreated automatically.
- If an add-on is updated through Apiserver, it will be reconfigured to the state given by the supplied fields in the initial config.
On the cluster, the add-ons are kept in /etc/kubernetes/addons on the master node, in
yaml / json files. The addon manager periodically kubectl applys the contents of this
directory. Any legit modification would be reflected on the API objects accordingly.
Particularly, rolling-update for deployments is now supported.
Each add-on must specify the following label: kubernetes.io/cluster-service: true.
Config files that do not define this label will be ignored. For those resources
exist in kube-system namespace but not in /etc/kubernetes/addons, addon manager
will attempt to remove them if they are attached with this label. Currently the other
usage of kubernetes.io/cluster-service is for kubectl cluster-info command to recognize
these cluster services.
The suggested naming for most types of resources is just <basename> (with no version
number) because we do not expect the resource name to change. But resources like Pod
, ReplicationController and DaemonSet are exceptional. As Pod updates may not change
fields other than containers[*].image or spec.activeDeadlineSeconds and may not add or
remove containers, it may not be sufficient during a major update. For ReplicationController,
most of the modifications would be legit, but the underlying pods would not got re-created
automatically. DaemonSet has similar problem as the ReplicationController. In these
cases, the suggested naming is <basename>-<version>. When version changes, the system will
delete the old one and create the new one (order not guaranteed).
Add-on update procedure
To update add-ons, just update the contents of /etc/kubernetes/addons
directory with the desired definition of add-ons. Then the system will take care
of:
- Removing objects from the API server whose manifest was removed.
- Creating objects from new manifests
- Updating objects whose fields are legally changed.
Cooperating with Horizontal / Vertical Auto-Scaling
As all cluster add-ons will be reconciled to the original state given by the initial config.
In order to make Horizontal / Vertical Auto-scaling functional, the related fields in config should
be left unset. More specifically, leave replicas in ReplicationController / Deployment
/ ReplicaSet unset for Horizontal Scaling, and leave resources for container unset for Vertical
Scaling. The periodical update won't include these specs, which will be managed by Horizontal / Vertical
Auto-scaler.