Automatic merge from submit-queue
invalid arg during applying SELinux label
Tested with Ubuntu 16.04 with SELINUX support. Actually, after `chcon `with long label got: `chcon: failed to set type security context component to ‘system_u:object_r:svirt_sandbox_file_t:s0’: Invalid argument`
Automatic merge from submit-queue
Use slices of items to clean up after tests
Fixes#27582.
We used to maintain a pointer variable for each process to kill after the
tests finish. @lavalamp suggested using a slice instead, which is a much
cleaner solution. This implements @lavalamp's suggestion and also extends
the idea to tracking directories that need to be removed after the tests finish.
This also means that we should no longer check for nil `killCmd`s inside
`func (k *killCmd) Kill() error {...}` (see #27582 and #27589). If a nil
`killCmd` makes it in there, something is bad elsewhere and we want to see
the nil pointer exception immediately.
Mentioning @timstclair and @euank wrt the original issue/PR.
Automatic merge from submit-queue
Should set default value for --service-node-port-range flag before verifying
For the flag `--service-node-port-range` of kube-apiserver, we know that it defaults to `30000-32767` if not specified. But if we only pass the flag `--kubernetes-service-node-port` with a valid value between `30000-32767` when starting kube-apiserver, a fatal error will occurs as the last below. It means that service port range is not 30000-32767 but empty. The log is from code [DefaultAndValidateRunOptions-->ValidateRunOptions-->verifyServiceNodePort](https://github.com/xiangpengzhao/kubernetes/blob/master/pkg/genericapiserver/genericapiserver.go#L580) where the flags are verified.
After tracing the apiserver related code, we can find the call stack:
```
func main() {
......
s := options.NewAPIServer()
......
app.Run(s)
......
}
```
In the `app.Run`, it calls [genericapiserver.DefaultAndValidateRunOptions(s.ServerRunOptions)](https://github.com/xiangpengzhao/kubernetes/blob/master/cmd/kube-apiserver/app/server.go#L80). But the `--kubernetes-service-node-port` hasn't been defaulted before there, so it's empty. It's then defaulted in `app.Run`-->[master.New](https://github.com/xiangpengzhao/kubernetes/blob/master/cmd/kube-apiserver/app/server.go#L276)-->[genericapiserver.New](https://github.com/xiangpengzhao/kubernetes/blob/master/pkg/master/master.go#L179)-->[setDefaults](https://github.com/xiangpengzhao/kubernetes/blob/master/pkg/genericapiserver/genericapiserver.go#L338)-->[defaultServiceNodePortRange](https://github.com/xiangpengzhao/kubernetes/blob/master/pkg/genericapiserver/genericapiserver.go#L281).
So, we have to set default value for `--kubernetes-service-node-port` in [NewServerRunOptions](https://github.com/xiangpengzhao/kubernetes/blob/master/pkg/genericapiserver/options/server_run_options.go#L105), as is done for `--secure-port` and/or `--insecure-port`. The `NewServerRunOptions` will be called in options.[NewAPIServer](https://github.com/xiangpengzhao/kubernetes/blob/master/cmd/kube-apiserver/app/options/options.go#L50)().
Hope that I have described the issue clearly. Thanks!
```
root@vm:~# kube-apiserver --etcd-servers=http://172.16.1.11:4001 --service-cluster-ip-range=192.168.122.0/24 --insecure-bind-address=0.0.0.0 --logtostderr=false --log-dir=/home/paas/zxp/log/kube --v=10 --kubernetes-service-node-port=30001 &
[2] 24629
root@vm:~# F0627 23:46:37.308726 24629 genericapiserver.go:580] Kubernetes service port range doesn't contain 30001
goroutine 1 [running]:
k8s.io/kubernetes/vendor/github.com/golang/glog.stacks(0x44f2500, 0x0, 0x0, 0x0)
/home/paas/zxp/code/k8s/fork/kubernetes/_output/local/go/src/k8s.io/kubernetes/vendor/github.com/golang/glog/glog.go:766 +0xb8
k8s.io/kubernetes/vendor/github.com/golang/glog.(*loggingT).output(0x44d2020, 0xc800000003, 0xc820238000, 0x438c73b, 0x13, 0x244, 0x0)
/home/paas/zxp/code/k8s/fork/kubernetes/_output/local/go/src/k8s.io/kubernetes/vendor/github.com/golang/glog/glog.go:717 +0x259
k8s.io/kubernetes/vendor/github.com/golang/glog.(*loggingT).printf(0x44d2020, 0xc800000003, 0x3223dc0, 0x33, 0xc8204c4cc8, 0x2, 0x2)
/home/paas/zxp/code/k8s/fork/kubernetes/_output/local/go/src/k8s.io/kubernetes/vendor/github.com/golang/glog/glog.go:655 +0x1d4
k8s.io/kubernetes/vendor/github.com/golang/glog.Fatalf(0x3223dc0, 0x33, 0xc8204c4cc8, 0x2, 0x2)
/home/paas/zxp/code/k8s/fork/kubernetes/_output/local/go/src/k8s.io/kubernetes/vendor/github.com/golang/glog/glog.go:1145 +0x5d
k8s.io/kubernetes/pkg/genericapiserver.verifyServiceNodePort(0xc8202a8400)
/home/paas/zxp/code/k8s/fork/kubernetes/_output/local/go/src/k8s.io/kubernetes/pkg/genericapiserver/genericapiserver.go:580 +0x1d5
k8s.io/kubernetes/pkg/genericapiserver.ValidateRunOptions(0xc8202a8400)
/home/paas/zxp/code/k8s/fork/kubernetes/_output/local/go/src/k8s.io/kubernetes/pkg/genericapiserver/genericapiserver.go:605 +0x2f
k8s.io/kubernetes/pkg/genericapiserver.DefaultAndValidateRunOptions(0xc8202a8400)
/home/paas/zxp/code/k8s/fork/kubernetes/_output/local/go/src/k8s.io/kubernetes/pkg/genericapiserver/genericapiserver.go:612 +0x4e
k8s.io/kubernetes/cmd/kube-apiserver/app.Run(0xc820224fc0, 0x0, 0x0)
/home/paas/zxp/code/k8s/fork/kubernetes/_output/local/go/src/k8s.io/kubernetes/cmd/kube-apiserver/app/server.go:80 +0x70
main.main()
/home/paas/zxp/code/k8s/fork/kubernetes/_output/local/go/src/k8s.io/kubernetes/cmd/kube-apiserver/apiserver.go:50 +0x121
[2]+ Exit 255 kube-apiserver --etcd-servers=http://172.16.1.11:4001 --service-cluster-ip-range=192.168.122.0/24 --insecure-bind-address=0.0.0.0 --logtostderr=false --log-dir=/home/paas/zxp/log/kube --v=10 --kubernetes-service-node-port=30001
root@vm:~#
```
Automatic merge from submit-queue
Add printf formatting for rrset objects.
Without this you just get two pointers in the debug log.
Before:
```
I0627 21:48:44.136615 1 dns.go:215] Existing recordset {0xc820168830 0xc820691540} is not equal to needed recordset &{0xc820168848 0xc820686040}, removing existing and adding needed.
```
After:
```
I0627 22:26:46.221856 1 dns.go:215] Existing recordset <(clouddns) "federated-service.e2e-tests-service-cuza5.federation.svc.us-central1-c.us-central1.kube.5yetis.net." type=CNAME rrdatas=["federated-service.e2e-tests-service-cuza5.federation.svc.us-central1.kube.5yetis.net."] ttl=180>
I0627 22:26:46.221885 1 dns.go:216] ... not equal to needed recordset <(clouddns) "federated-service.e2e-tests-service-cuza5.federation.svc.us-central1-c.us-central1.kube.5yetis.net." type=CNAME rrdatas=["federated-service.e2e-tests-service-cuza5.federation.svc.us-central1.kube.5yetis.net."] ttl=180>
I0627 22:26:46.221919 1 dns.go:217] ... removing existing and adding needed.
```
Without this you just get two pointers in the debug log.
Before:
I0627 21:48:44.136615 1 dns.go:215] Existing recordset {0xc820168830 0xc820691540} is not equal to needed recordset &{0xc820168848 0xc820686040}, removing existing and adding needed.
After:
I0627 22:26:46.221856 1 dns.go:215] Existing recordset <(clouddns) "federated-service.e2e-tests-service-cuza5.federation.svc.us-central1-c.us-central1.kube.5yetis.net." type=CNAME rrdatas=["federated-service.e2e-tests-service-cuza5.federation.svc.us-central1.kube.5yetis.net."] ttl=180>
I0627 22:26:46.221885 1 dns.go:216] ... not equal to needed recordset <(clouddns) "federated-service.e2e-tests-service-cuza5.federation.svc.us-central1-c.us-central1.kube.5yetis.net." type=CNAME rrdatas=["federated-service.e2e-tests-service-cuza5.federation.svc.us-central1.kube.5yetis.net."] ttl=180>
I0627 22:26:46.221919 1 dns.go:217] ... removing existing and adding needed.
Automatic merge from submit-queue
Implement custom help command for kubectl
```release-note
* kubectl help now provides "Did you mean this?" suggestions for typo/invalid command names.
```
Custom implementation of help command allows to print `Did you mean this?` with
suggestions, which is missed in embed help command from github.com/spf13/cobra
Also, it can be extended with different search features. At this patch, help
command searches query in short descriptions of commands in case of mismatch
with commands names.
fixes#25234
Automatic merge from submit-queue
Federated Services e2e: Simplify logic and logging around verificatio…
Simplify logic and logging around verification of underlying services.
Fixes#28269.
Without this PR, service verification in 4 of our e2e tests sometimes fails.
Own implemenation of help command allows to print `Did you mean this?` with
suggestions, which is missed in embed help command from github.com/spf13/cobra
Also, it can be extended with different search features. At this patch, help
command search query in short descriptions of commands in case of mismatch
with commands names.
fixes#25234
Automatic merge from submit-queue
oidc auth plugin: don't hard fail if provider is unavailable
When using OpenID Connect authentication, don't cause the API
server to fail if the provider is unavailable. This allows
installations to run OpenID Connect providers after starting the
API server, a common case when the provider is running on the
cluster itself.
Errors are now deferred to the authenticate method.
cc @sym3tri @erictune @aaronlevy @kubernetes/sig-auth
Automatic merge from submit-queue
[kubelet] Allow opting out of automatic cloud provider detection in kubelet. By default kubelet will auto-detect cloud providers
fixes#28231
Automatic merge from submit-queue
Fix problems with container restarts and flocker volumes
[]()
The issue is described in #22436. A pod with a flocker volume attached can't be restarted. A complete new pod has to be created, then everything is working fine.
The problem is around the `isReady` check. This ready state is not getting cleaned up for restarting containers. IMHO we don't need this ready state stored at all. Maybe it was meant to not call the flocker API during pod restarts, but I would rather have it removed... (happy about some hints why we need this lines)
I tested it in a AWS/flocker cluster: starting/restarting/rescheduling on another node (in the same zone) works perfectly fine.
Automatic merge from submit-queue
Bump skydns godeps to latest
Update Godeps for github.com/skynetservices/skydns and miekg/dns.
Bump kubedns version to 1.6 with latest skynetservices/skydns code
Built kube-dns for all architectures and pushed containers to gcr.io.
When using OpenID Connect authentication, don't cause the API
server to fail if the provider is unavailable. This allows
installations to run OpenID Connect providers after starting the
API server, a common case when the provider is running on the
cluster itself.
Errors are now deferred to the authenticate method.
Automatic merge from submit-queue
Lock all possible kubecfg files at the beginning of ModifyConfig.
Prevent concurrent calls to ModifyConfig on the same (or overlapping) kubeconfig files.
Automatic merge from submit-queue
Add error log for Run function in server.go
When clientcmd.BuildConfigFromFlags and os.Hostname return error, there are no log information in Run function, neither did the upper function(main of scheduler), add it, I suggest.
Automatic merge from submit-queue
Ignore cmd/libs/go2idl/generator when running coverage
Temporary fix so we can run coverage (with a PASS).
fixes#24967
Automatic merge from submit-queue
Fixed goroutinemap race on Wait()
sync.WaitGroup produces data races when a GoroutineMap is empty and Wait() and Run() are called at the same time.
From sync.WaitGroup:
> Note that calls with a positive delta that occur when the counter is zero must happen before a Wait.
Fixes#28128
Note that this issue affects only PersistentVolume unit tests.
@saad-ali, PTAL
Fixes#27582
We used to maintain a pointer variable for each process to kill after the
tests finish. @lavalamp suggested using a slice instead, which is a much
cleaner solution. This implements @lavalamp's suggestion and also extends
the idea to tracking directories that need to be removed after the tests finish.
This also means that we should no longer check for nil `killCmd`s inside
`func (k *killCmd) Kill() error {...}` (see #27582 and #27589). If a nil
`killCmd` makes it in there, something is bad elsewhere and we want to see
the nil pointer exception immediately.
Automatic merge from submit-queue
Enhance kubedns pod health checks to cover kubedns container
The existing health check hits port 53, the dnsmasq container, with the same domain name every time. Since dnsmasq looks up and caches results from the kubedns container, running on port 10053, the health check is not covering the kubedns container after the first query (and once every TTL expiration).
This PR enhances the health check to directly hit port 10053 (kubedns) in addition to port 53.
