Got the proxy-server coming up in the master.
Added certs and have it comiung up with those certs.
Added a daemonset to run the network-agent.
Adding support for agent running as a sameon set on every node.
Added quick hack to test that proxy server/agent were correctly
tunneling traffic to the kubelet.
Added more WIP for reading network proxy configuration.
Get flags set correctly and fix connection services.
Adding missing ApplyTo
Added ConnectivityService.
Fixed build directives. Added connectivity service configuration.
Fixed log levels.
Fixed minor issues for feature turned off.
Fixed boilerplate and format.
Moved log dialer initialization earlier as per Liggits suggestion.
Fixed a few minor issues in the configuration for GCE.
Fixed scheme allocation
Adding unit test.
Added test for direct connectivity service.
Switching to injecting the Lookup method rather than using a Singleton.
First round of mikedaneses feedback.
Fixed deployment to use yaml and other changes suggested by MikeDanese.
Switched network proxy server/agent which are kebab-case not camelCase.
Picked up DIAL_RSP fix.
Factored in deads2k feedback.
Feedback from mikedanese
Factored in second round of feedback from David.
Fix path in verify.
Factored in anfernee's feedback.
First part of lavalamps feedback.
Factored in more changes from lavalamp and mikedanese.
Renamed network-proxy to konnectivity-server and konnectivity-agent.
Fixed tolerations and config file checking.
Added missing strptr
Finished lavalamps requested rename.
Disambiguating konnectivity service by renaming it egress selector.
Switched feature flag to KUBE_ENABLE_EGRESS_VIA_KONNECTIVITY_SERVICE
This commit adds support for using `gke-exec-auth-plugin` (vTPM-based
certificates for mTLS) for webhooks when calling endpoints matching
`*.googleapis.com`, and integrates this support with
ValidatingAdmissionWebhook.
To enable it, request ValidatingAdmissionWebhook with
`ADMISSION_CONTROL=...,ValidatingAdmissionWebhook,...` (default) and
opt in to `gke-exec-auth-plugin` using `WEBHOOK_GKE_EXEC_AUTH=true`
during the configuration process.
If you don't opt-in, ValidatingAdmissionWebhook will be deployed as
before.
Requesting `WEBHOOK_GKE_EXEC_AUTH=true` will fail if you have not
provided other configuration variables:
* `EXEC_AUTH_PLUGIN_URL`: controls whether `gke-exec-auth-plugin` is
downloaded during the installation step. A prerequisite for
actually using the plugin.
* `TOKEN_URL`, `TOKEN_BODY`, and `TOKEN_BODY_UNQUOTED`:
configuration values used when calling the plugin. `TOKEN_URL`
and `TOKEN_BODY` have existing usage. `TOKEN_BODY_UNQUOTED` is a
new variable that is meant to sidestep the problem of inverting
`strconv.Quote` in Bash.
The existing configuration process for ImagePolicyWebhook has been
reworked to make it play nicely with ValidatingAdmissionWebhook under
`WEBHOOK_GKE_EXEC_AUTH=true`.
* It originally placed the ImagePolicyWebhook configuration object
at the top-level of the file specified by
`--admission-control-config-file`. I can't see why this worked;
it must have been hitting some sort of lucky path through the
various config file loading mechanisms. Now, it places its
configuration in a sub-field of that file, which is shared among
all admission control plugins.
* It mounted its various config files read-write. I reviewed the
code and couldn't see why it was necessary, so I moved the config
files into the existing read-only mount at `/etc/srv/kubernetes`.
* It now checks that all the configuration values it requires have
been provided.
Co-authored-by: Mike Danese <mikedanese@google.com>
Co-authored-by: Taahir Ahmed <taahm@google.com>
using `local -r` will blow up, example output:
```
/home/kubernetes/bin/configure.sh: line 388: local: manifest_name: readonly variable
```
Change-Id: Id379180803d44dd9c7ac0da41c1cd56de0fe54a4
The feature caused tests to fail when it was enabled.
- https://github.com/kubernetes/kubernetes/issues/78628
Work is in progress to fix the feature, but until that work is complete,
we will disable it in the GCE scripts.
Split arguments to be passed to cluster autoscaler binary,
so each argument is passed separately.
This is preparatory work for migrating CA to disroless base image
and passing multiple arguments together does not work if CA is
not wrapped around with shell script
Change-Id: I26b5a764d2a12079c7f4ed6633ccabf8d623e232
Till a few days ago, it was possible to ssh into master and access cluster via insecure master port.
Now, the master insecure port has been disabled, we're not able to do that anymore.
This PR aims to fix that by uploading the kubeconfig to the master metadata during cluster setup in tests.
* Touched containers: kube-apiserver, kube-scheduler,
kube-controller-manager.
* Remove the shell dependencies when upstart the containers.
* Reformat the command parameters to ["Exec", "Param1", "Param2"]
This change bumps the stackdriver logging agent version to v1-9,
re-enable it, and change the script/configuration to:
* Create /var/log in the startup script, since the fluentd
configuration expects the directory to exists
* Add support for collecting kubelet/kube-proxy logs
* Touched containers: kube-apiserver, kube-scheduler,
kube-controller-manager.
* Remove the shell dependencies when upstart the containers.
* Reformat the command parameters to ["Exec", "Param1", "Param2"]
- Pass most of the flags via kube-env, similar to Kubelet, to make it
easier to reconfigure.
- Use kube-proxy's kubeconfig file to pass the API server address,
rather than the --master flag.
- Make KUBEPROXY_ARGS a required variable in Windows kube-env, similar
to KUBELET_ARGS.
I recently lowered the timeout value to 2 minutes but this turned out to
be too aggressive for now. 3 minutes is usually long enough for the test
containers to come up.
Note that this test is only used during development, it does not run
continuously anywhere.
--hostname-override should not be necessary, we can rely on kube-proxy
to auto-detect it just like we do in Kubelet.
I also added an optional variable to kube_env, KUBEPROXY_ARGS, that we
can use to remotely extend kube-proxy's command-line if necessary.
This PR adds the support for GCEPD volume. For now, it uses a workaround
(https://github.com/pjh/gce-tools) to get disk number in windows for a given GCE PD name.
Also switches to the most recent 64-bit version of OpenSSH for Windows.
Tested:
PROJECT=${CLOUDSDK_CORE_PROJECT} KUBERNETES_SKIP_CONFIRM=y NUM_NODES=2 \
NUM_WINDOWS_NODES=2 KUBE_GCE_ENABLE_IP_ALIASES=true TEST_CLUSTER=true \
./cluster/kube-up.sh
This rule is only created if the cluster has one or more Windows nodes,
but delete-firewall-rules() ignores firewall rule arguments that do not
exist so it's safe to always attempt to delete this rule.
Looks like the get-master-root-disk-size() and get-master-disk-size() functions didn't anticipate
clusters bigger than 2K nodes.
In https://github.com/kubernetes/kubernetes/issues/72976 we found out
that 100GB may be not enough for large clusters (5K nodes) when it comes
to master root disk size.
Updating both get-master-root-disk-size() and get-master-disk-size() to
make them consistent and match cluster sizes with get-master-size()
function.
Currently, only kubernetes-server-linux-<arch>.tar.gz will be downloaded
and staged for tests. It is not possible to test a cluster where the
platform/arch of the nodes differs from that of the master
control-plane. This changes allows downloading and staging node
binaries, in addition to the existing server binaries.
This change renames the '--experimental-encryption-provider-config'
flag to '--encryption-provider-config'. The old flag is accepted but
generates a warning.
In 1.14, we will drop support for '--experimental-encryption-provider-config'
entirely.
Co-authored-by: Stanislav Laznicka <slaznick@redhat.com>
This change includes the yaml files and gce startup script changes
to run this addon. It is disabled by default, can be enabled by setting
KUBE_ENABLE_NODELOCAL_DNS=true
An ip address is required for the cache instance to listen for
requests on, default is a link local ip address of value 169.254.25.10
addressed review comments, updated image location
Picked a different prometheus port so stats port is not same as the
coredns deployment
Removed the nodelocaldns-ready label.
Set memory limit to 30Mi
2. Do not set Preemptible node label on master node unless the master node is explicitly set to be preemptible.
Signed-off-by: Vishnu kannan <vishnuk@google.com>
Automatic merge from submit-queue (batch tested with PRs 67950, 68195). If you want to cherry-pick this change to another branch, please follow the instructions here: https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md.
Remove e2e-image-puller
**What this PR does / why we need it**:
A long time ago, We added the image prepulling as a workaround due to
the overwhelming amount of flake caused by pulling during the tests.
This functionality has been broken for a while now when we switched to a
COS image where mounting `docker` binary into `busybox` stopped working.
So we just have dead code we should clean up.
Change-Id: I538171a5c1d9361eee7f9e0a99655b88b1721e3e
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Fixes#63355
**Special notes for your reviewer**:
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions here: https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md.
Update default etcd server to 3.2.24 for kubernetes 1.12
**What this PR does / why we need it**:
Update default etcd server to 3.2.24 for kubernetes 1.12
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
xref #68147
**Special notes for your reviewer**:
NONE
**Release note**:
```
Update default etcd server to 3.2.24 for kubernetes 1.12
```
/assign @wojtek-t @jpbetz @dims
/cc @kubernetes/sig-cluster-lifecycle-pr-reviews @gyuho
Automatic merge from submit-queue (batch tested with PRs 68119, 68191). If you want to cherry-pick this change to another branch, please follow the instructions here: https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md.
Update cluster autoscaler to 1.12.0-beta.1
Update Cluster Autoscaler to version 1.12.0-beta.1 which is compatible with k8s 1.12.
Note: this is pre release version. Update to the final version of CA image will be done a week before k8s release deadline.
Version skip from 1.3.x to 1.12.x is to synchronize version numbering between Cluster Autoscaler and k8s core.
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 68087, 68256, 64621, 68299, 68296). If you want to cherry-pick this change to another branch, please follow the instructions here: https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md.
Bump addon-manager to v8.7
**What this PR does / why we need it**:
Major changes:
- Support extra `--prune-whitelist` resources in kube-addon-manager.
- Update kubectl to v1.10.7.
Basically picking up https://github.com/kubernetes/kubernetes/pull/67743.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Fixes #NONE
**Special notes for your reviewer**:
/assign @Random-Liu @mikedanese
**Release note**:
```release-note
Bump addon-manager to v8.7
- Support extra `--prune-whitelist` resources in kube-addon-manager.
- Update kubectl to v1.10.7.
```
Automatic merge from submit-queue (batch tested with PRs 68087, 68256, 64621, 68299, 68296). If you want to cherry-pick this change to another branch, please follow the instructions here: https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md.
gce: use getrandom instead of urandom for on node rng
```release-note
NONE
```
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions here: https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md.
Fix an issue about generation of secret key with invalid key size in gce
Replace /dev/random to /dev/urandom to avoid generation of secret key with invalid key size.
**What this PR does / why we need it**:
In GCE env, AESGCM encryption of secrets by default generates a secret key with /dev/random which sometime generates a key with invalid size.
This cause the cluster/kube-up.sh to fail in gce environment.
This PR replaces /dev/random with /dev/urandom to have a secret key generated consistently with right size.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Fixes#67091
**Special notes for your reviewer**:
**Release note**:
```release-note
cluster/gce: generate consistent key sizes in config-default.sh using /dev/urandom instead of /dev/random
```
Automatic merge from submit-queue (batch tested with PRs 68161, 68023, 67909, 67955, 67731). If you want to cherry-pick this change to another branch, please follow the instructions here: https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md.
Register RuntimeClass CRD as an addon
**What this PR does / why we need it**:
Register the RuntimeClass CRD when the RuntimeClass feature gate is enabled. This is done in through the addon manager.
This is an alternative approach to https://github.com/kubernetes/kubernetes/pull/67924
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
For https://github.com/kubernetes/features/issues/585
**Release note**:
Covered by #67737
```release-note
NONE
```
/sig node
/kind feature
/priority important-soon
/milestone v1.12
In the context, our urandoms where generally safe, however getrandom has
built in invariants around entropy pool initialization, making getrandom
safe in all contexts. This should protect us from cryptopasta errors or
weird entropy issues.
Automatic merge from submit-queue (batch tested with PRs 67736, 68123, 68138). If you want to cherry-pick this change to another branch, please follow the instructions here: https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md.
Added support to get clusters in gce cloud provider.
**What this PR does / why we need it**:
Implemented the call to get all cluster objects in a zone for a project.
Also added code to allow the container api to be set in the gce.conf
file.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Fixes #
**Special notes for your reviewer**:
**Release note**:
```release-note
NONE
```
A long time ago, We added the image prepulling as a workaround due to
the overwhelming amount of flake caused by pulling during the tests.
This functionality has been broken for a while now when we switched to a
COS image where mounting `docker` binary into `busybox` stopped working.
So we just have dead code we should clean up.
Change-Id: I538171a5c1d9361eee7f9e0a99655b88b1721e3e
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions here: https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md.
Support extra prune resources in kube-addon-manager.
The default prune whitelist resources in https://github.com/kubernetes/kubernetes/blob/master/pkg/kubectl/cmd/apply.go#L531 are sometimes not enough.
One example is that when we remove an admission webhook running as an addon pod, after we remove the addon yaml file, the admission webhook pod will be pruned, but the `MutatingWebhookConfiguration`/`ValidationWebhookConfiguration` won't... If the webhook failure policy is `Fail`, this will break the cluster, and users can't create new pods anymore.
It would be good to at least make this configurable, so that users and vendors can configure it based on their requirement.
This PR keeps the default prune resource list exactly the same with before, just makes it possible to add extra ones.
@dchen1107 @MrHohn @kubernetes/sig-cluster-lifecycle-pr-reviews @kubernetes/sig-gcp-pr-reviews
Signed-off-by: Lantao Liu <lantaol@google.com>
**Release note**:
```release-note
Support extra `--prune-whitelist` resources in kube-addon-manager.
```
Automatic merge from submit-queue (batch tested with PRs 65251, 67255, 67224, 67297, 68105). If you want to cherry-pick this change to another branch, please follow the instructions here: https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md.
Add namespace for (cluster)role(binding) cloud-provider.
**What this PR does / why we need it**:
Add namespace for (cluster)role(binding) cloud-provider.
Change the addonmanager mode to be from reconcile to EnsureExists.
Needs to be cherrypicked together with https://github.com/kubernetes/kubernetes/pull/59686.
**Special notes for your reviewer**:
/assign @bowei @tallclair
/sig auth
**Release note**:
```release-note
Role, ClusterRole and their bindings for cloud-provider is put under system namespace. Their addonmanager mode switches to EnsureExists.
```
Manual tested. Cluster can be created succesfully using kube-up.sh with desired (cluster)role(binding)s.
Automatic merge from submit-queue (batch tested with PRs 64283, 67910, 67803, 68100). If you want to cherry-pick this change to another branch, please follow the instructions here: https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md.
Adding GCE node termination handler as an optional addon.
This step is a pre-requisite for auto-deploying that addon in GKE
cc @mikedanese
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions here: https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md.
Promote AdvancedAuditing to GA
**What this PR does / why we need it**:
Removes deprecated legacy code used for basic audit logging in favor of advanced audit logging.
```release-note
Promote AdvancedAuditing to GA, replacing the previous (legacy) audit logging mechanisms.
```
Automatic merge from submit-queue (batch tested with PRs 67745, 67432, 67569, 67825, 67943). If you want to cherry-pick this change to another branch, please follow the instructions here: https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md.
Add flag for disabling prometheus-to-sd only for daemon sets
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 67745, 67432, 67569, 67825, 67943). If you want to cherry-pick this change to another branch, please follow the instructions here: https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md.
Enable CoreDNS as default for kube-up deployments
**What this PR does / why we need it**:
Enable CoreDNS as default (for kube-up installations)
It will allow to run CI tests to prepare graduation criteria for CoreDNS as Default
see : KEP - https://github.com/kubernetes/community/blob/master/keps/sig-network/0012-20180518-coredns-default-proposal.md
see also : https://github.com/kubernetes/features/issues/566
NOTE for release : I guess that CoreDNS as default server for k8s needs a longer description. This specific PR is to ensure we validate all e2e.
```release-note
Make CoreDNS be the default DNS server in kube-up (instead of kube-dns formerly).
It is still possible to deploy kube-dns by setting CLUSTER_DNS_CORE_DNS=false.
```
The requested Service Protocol is checked against the supported protocols of GCE Internal LB. The supported protocols are TCP and UDP.
SCTP is not supported by OpenStack LBaaS. If SCTP is requested in a Service with type=LoadBalancer, the request is rejected. Comment style is also corrected.
SCTP is not allowed for LoadBalancer Service and for HostPort. Kube-proxy can be configured not to start listening on the host port for SCTP: see the new SCTPUserSpaceNode parameter
changed the vendor github.com/nokia/sctp to github.com/ishidawataru/sctp. I.e. from now on we use the upstream version.
netexec.go compilation fixed. Various test cases fixed
SCTP related conformance tests removed. Netexec's pod definition and Dockerfile are updated to expose the new SCTP port(8082)
SCTP related e2e test cases are removed as the e2e test systems do not support SCTP
sctp related firewall config is removed from cluster/gce/util.sh. Variable name sctp_addr is corrected to sctpAddr in pkg/proxy/ipvs/proxier.go
cluster/gce/util.sh is copied from master
Implemented the call to get all cluster objects in a zone for a project.
Also added code to allow the container api to be set in the gce.conf
file.
Requested fix for @lavalamp. Fixed GetClusters to be GetManagedClusters.
Leaving ListClusters as ListClusters as it is part of the Cloud Clusters
interface, despite also being a "managed" call.
Remove copy pasta :D
Fixed method variable name.
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
remove rescheduler since scheduling DS pods by default scheduler is moving to beta
**What this PR does / why we need it**:
remove rescheduler since scheduling DS pods by default scheduler is moving to beta
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Fixes#64725
**Special notes for your reviewer**:
**Release note**:
```release-note
Remove rescheduler since scheduling DS pods by default scheduler is moving to beta.
```
Automatic merge from submit-queue. If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Set --audit-webhook-mode flag properly based on the env variable
**What this PR does / why we need it**:
Audit webhook mode control via env variable is currently broken and this is the fix.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
**Special notes for your reviewer**:
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 66177, 66185, 67136, 67157, 65065). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Update configure-helper.sh to support heapster resource optimizations
**What this PR does / why we need it**:
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Fixes #
**Special notes for your reviewer**:
**Release note**:
```release-note
NONE
```
Automatic merge from submit-queue (batch tested with PRs 66177, 66185, 67136, 67157, 65065). If you want to cherry-pick this change to another branch, please follow the instructions <a href="https://github.com/kubernetes/community/blob/master/contributors/devel/cherry-picks.md">here</a>.
Templatize the scaling policy for metrics-server
**What this PR does / why we need it**:
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Fixes #
**Special notes for your reviewer**:
**Release note**:
```release-note
NONE
```
