Automatic merge from submit-queue
Don't double encode protobuf runtime.Unknown
When using runtime.Object and runtime.RawExtension, passing
runtime.Unknown to protobuf serializer should result in the raw message
being serialized (since all normal protobuf objects are
runtime.Unknown).
Also, avoid setting a content-type when decoding a protobuf object
except when the content appears to be proto.
@wojtek-t
Automatic merge from submit-queue
Generate a better Stringer method for proto types
This replaces the bad string output generated by golang/proto with gogo/protobuf stringer generation. Makes the output similar to %#v and more debuggable. We have to have a String() method to implement proto.Message, so this is strictly better.
@wojtek-t, @thockin for after your PR merges
Fixes#28756
Automatic merge from submit-queue
Implement a RESTMappings method
With the introduction of batch/v1 and batch/v2alpha1, any
MultiRESTMapper that has per groupversion mappers (as a naive discovery
client would create) would end up being unable to call RESTMapping() on
batch.Jobs. As we finish up discovery we will need to be able to choose
prioritized RESTMappings based on the service discovery doc.
This change implements RESTMappings(groupversion) which returns all
possible RESTMappings for that kind. That allows a higher level call to
prioritize the returned mappings by server or client preferred version.
@deads2k
Automatic merge from submit-queue
[GarbageCollector] Let the RC manager set/remove ControllerRef
What's done:
* RC manager sets Controller Ref when creating new pods
* RC manager sets Controller Ref when adopting pods with matching labels but having no controller
* RC manager clears Controller Ref when pod labels change
* RC manager clears pods' Controller Ref when rc's selector changes
* RC manager stops adoption/creating/deleting pods when rc's DeletionTimestamp is set
* RC manager bumps up ObservedGeneration: The [original code](https://github.com/kubernetes/kubernetes/blob/master/pkg/controller/replication/replication_controller_utils.go#L36) will do this.
* Integration tests:
* verifies that changing RC's selector or Pod's Labels triggers adoption/abandoning
* e2e tests (separated to #27151):
* verifies GC deletes the pods created by RC if DeleteOptions.OrphanDependents=false, and orphans the pods if DeleteOptions.OrphanDependents=true.
TODO:
- [x] we need to be able to select Pods that have a specific ControllerRef. Then each time we sync the RC, we will iterate through all the Pods that has a controllerRef pointing the RC, event if the labels of the Pod doesn't match the selector of RC anymore. This will prevent a Pod from stuck with a stale controllerRef, which could be caused by the race between abandoner (the goroutine that removes controllerRef) and worker the goroutine that add controllerRef to pods).
- [ ] use controllerRef instead of calling `getPodController`. This might be carried out by the control-plane team.
- [ ] according to the controllerRef proposal (#25256): "For debugging purposes we want to add an adoptionTime annotation prefixed with kubernetes.io/ which will keep the time of last controller ownership transfer." This might be carried out by the control-plane team.
cc @lavalamp @gmarek
Automatic merge from submit-queue
Modify wrong description about "InsecurePort" in genericapiserver.go
Modify wrong description about "InsecurePort" in genericapiserver.go, avoid misleading.
Automatic merge from submit-queue
Improve fatal error description for init in genericapiserver.go
When api.NewRequestContextFilter return error in the "init" function of genericapiserver.go, there are no handler info, add more information to indicate s.Handler or s.InsecureHandler, I suggest.
Automatic merge from submit-queue
Optimise the New function in genericapiserver.go
It needn't the handlerContainer variable in the "New" function when deal s.HandlerContainer, I think, and it seems the codes more concise.
Automatic merge from submit-queue
[garbage collector] add e2e test
This PR also includes some changes to plumb controller-manager's `--enable_garbage_collector` from the environment variable.
The e2e test will not be run by the core suite because it's marked `[Feature:GarbageCollector]`.
The corresponding jenkins job configuration PR is https://github.com/kubernetes/test-infra/pull/132.
Move SystemReserved and KubeReserved into KubeletConfiguration struct
Convert int64 to int32 for some external type fields so they match internal ones
tLS* to tls* for JSON field names
Fix dependency on removed options.AutoDetectCloudProvider
Change floats in KubeletConfiguration API to ints
Update external KubeletConfiguration type
Add defaults for new KubeletConfiguration fields
Modify some defaults to match upstream settings
Add/rename some conversion functions
Updated codegen
Fixed typos
Mike Danese caught that s.NodeLabels wasn't allocated, fix on line 118
of cmd/kubelet/app/options/options.go.
Provide list of valid sources in comment for HostNetworkSources field
Automatic merge from submit-queue
Fix watch cache filtering
When serving watch events for a particular namespace, the watch cache filters out events from other namespaces by checking the etcd key of the event's object, and making sure it is prefixed with the root key for the namespace being watched.
The prefix check does not ensure the match occurs on a path segment boundary, so a watch on namespace `test` is delivered watch events for namespace `test1`.
This tightens the check to ensure the prefix match occurs on a path segment boundary.
When using runtime.Object and runtime.RawExtension, passing
runtime.Unknown to protobuf serializer should result in the raw message
being serialized (since all normal protobuf objects are
runtime.Unknown).
Also, avoid setting a content-type when decoding a protobuf object
except when the content appears to be proto.
Automatic merge from submit-queue
Support terminal resizing for exec/attach/run
```release-note
Add support for terminal resizing for exec, attach, and run. Note that for Docker, exec sessions
inherit the environment from the primary process, so if the container was created with tty=false,
that means the exec session's TERM variable will default to "dumb". Users can override this by
setting TERM=xterm (or whatever is appropriate) to get the correct "smart" terminal behavior.
```
Fixes#13585
Automatic merge from submit-queue
Log restclient request/response bodies using %#v.
These are now protobufs, so outputting them with %s dumps a large
amount of binary garbage into the log. %#v properly escapes exotic
characters.
Automatic merge from submit-queue
Retry when apiserver fails to listen on insecure port
The apiserver will already continually retry when it fails to bind to the secure port. However, with the insecure port it does not retry, and any failures cause the apiserver to exit.
This change makes it so the api-server will retry on both insecure/secure ports.
A use-case for this change is for self-hosting the api-server, particularly when you are only running a single copy in your cluster. In some bootstrap and upgrade scenarios - it's necessary to replace/pivot from an existing api-server on the same host -- where you need a new copy running before tearing down the old (which is problematic when the api-server will try to bind to an in-use port and exit).
With the introduction of batch/v1 and batch/v2alpha1, any
MultiRESTMapper that has per groupversion mappers (as a naive discovery
client would create) would end up being unable to call RESTMapping() on
batch.Jobs. As we finish up discovery we will need to be able to choose
prioritized RESTMappings based on the service discovery doc.
This change implements RESTMappings(groupversion) which returns all
possible RESTMappings for that kind. That allows a higher level call to
prioritize the returned mappings by server or client preferred version.
Automatic merge from submit-queue
update resource builder error message to be more clear
release-note-none
The error message given by command line `kubectl get` is sometimes of no help / not clear on what must be corrected, e.g.：
`kubectl get pod pod/database-1-i10b9`
error: when passing arguments in resource/name form, all arguments must include the resource
##### Steps to Reproduce:
1. Run command "$ kubectl get pod pod/database-1-i10b9"
##### Actual Result:
Get unfriendly error message which is of no help:
"error: when passing arguments in resource/name form, all arguments must include the resource"
##### Expected Result:
Error message should recommend end user to run this cli in good grammar: "$ kubectl get pod database-1-i10b9" or "$ kubectl get pod/database-1-i10b9"
##### Before
"error: when passing arguments in resource/name form, all arguments must include the resource"
##### After
"error: there is no need to specify a resource type as a separate argument when passing arguments in resource/name form (e.g. `kubectl get resource/<resource_name>` instead of `kubectl get resource resource/<resource_name>`"
[]()
Add support for terminal resizing for exec, attach, and run. Note that for Docker, exec sessions
inherit the environment from the primary process, so if the container was created with tty=false,
that means the exec session's TERM variable will default to "dumb". Users can override this by
setting TERM=xterm (or whatever is appropriate) to get the correct "smart" terminal behavior.
