It makes more sense to use advertised etcd endpoints instead of using
listened client URLS.
I'm solving this use case: Listen on localhost and public IP but I don't
want to add multiple endpoints to kube-apiservers because it's targeting
same etcd server.
Replace the unnecessary use of InitConfiguration in images.go with
ClusterConfiguration. This changes the interfaces of the following functions:
- GetKubeControlPlaneImage
- GetEtcdImage
- GetAllImages
Signed-off-by: Rostislav M. Georgiev <rostislavg@vmware.com>
Automatic merge from submit-queue (batch tested with PRs 64593, 65117, 65629, 65827, 65686). 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>.
[kubeadm] Remove usage of `PersistentVolumeLabel` in phases/controlplane/manifests*.go
**What this PR does / why we need it**:
Remove usage of `PersistentVolumeLabel` in phases/controlplane/manifests*.go
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Close [kubernetes/kubeadm#866](https://github.com/kubernetes/kubeadm/issues/866)
**Special notes for your reviewer**:
/assign @neolit123
**Release note**:
```release-note
kubeadm: Remove usage of `PersistentVolumeLabel`
```
GetCoreImage is a too generic function, that takes too many arguments. This
makes it prone to errors that may be difficult to trace. The solution is to
split it into the following couple of functions with a more targeted interface:
- GetKubeControlPlaneImage used to fetch Kubernetes control plane images or the
unified control plane image (if one is specified).
- GetEtcdImage is used to fetch the etcd image.
In addition to these, a couple of new utility functions are also created:
- GetKubeControlPlaneImageNoOverride used like GetKubeControlPlaneImage but
does not return the unified control plane image (even if it is set).
- GetGenericArchImage returns image path in the form of "prefix/image-goarch:tag"
Signed-off-by: Rostislav M. Georgiev <rostislavg@vmware.com>
This follows the pattern `kubectl` uses for logging.
There are two remaining glog.Infof call that cannot be removed easily.
One glog call comes from kubelet validation which calls features.SetFromMap.
The other comes from test/e2e during kernel validation.
Mostly fixeskubernetes/kubeadm#852
Signed-off-by: Chuck Ha <ha.chuck@gmail.com>
The API server argument --admission-control is deprecated.
Use the following arguments instead:
--enable-admission-plugins=NodeRestriction
--disable-admission-plugins=PersistentVolumeLabel
Add comment that PersistentVolumeLabel should be removed at some
point in 1.11.
kubeadm uses LeaseEndpointReconcilerType as import from
k8s.io/kubernetes/pkg/master/reconcilers. However, this pull a huge
load of extra dependencies (among which
pkg/client/clientset_generated/internalclientset). The solution is
to copy this string constant locally in kubeadm.
Signed-off-by: Rostislav M. Georgiev <rostislavg@vmware.com>
Automatic merge from submit-queue (batch tested with PRs 63735, 62807). 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>.
Kubeadm: Update comments and UT to remove /66 restriction
**What this PR does / why we need it**:
The comments in cmd/kubeadm/app/phases/controlplane/manifests.go mention the
IPv6 /66 restriction, and the UT also refers to this.
This restriction was removed in PR#60089
This removes those comments and updates the UT
**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#62806
**Special notes for your reviewer**:
**Release note**:
```release-note-none
```
The comments in cmd/kubeadm/app/phases/controlplane/manifests.go mention the
IPv6 /66 restriction, and the UT also refers to this.
This restriction was removed in PR#60089
Removes an unused parameter in getAPIServerCommand
Cleans up tests by:
* Naming the tests
* Using t.Run for better test output
* Removing duplicates
Fixeskubernets/kubeadm#760
Signed-off-by: Chuck Ha <ha.chuck@gmail.com>
Automatic merge from submit-queue (batch tested with PRs 59027, 62333, 57661, 62086, 61584). 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>.
Implement verbosity feature for kubeadm
[WIP] Adds verbosity feature to init command hierarchy of kubeadm utility.
**What this PR does / why we need it**:
Implements verbosity feature to kubeadm
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Fixeskubernetes/kubeadm#340
**Special notes for your reviewer**:
I will be splitting this work into a smaller PR to keep it separate and clean.
**Release note**:
```release-note
Implements verbosity logging feature for kubeadm commands
```
- Place etcd server and peer certs & keys into pki subdir
- Move certs.altName functions to pkiutil + add appendSANstoAltNames()
Share the append logic for the getAltName functions as suggested by
@jamiehannaford.
Move functions/tests to certs/pkiutil as suggested by @luxas.
Update Bazel BUILD deps
- Warn when an APIServerCertSANs or EtcdCertSANs entry is unusable
- Add MasterConfiguration.EtcdPeerCertSANs
- Move EtcdServerCertSANs and EtcdPeerCertSANs under MasterConfiguration.Etcd
- Generate Server and Peer cert for etcd
- Generate Client cert for apiserver
- Add flags / hostMounts for etcd static pod
- Add flags / hostMounts for apiserver static pod
- Generate certs on upgrade of static-pods for etcd/kube-apiserver
- Modify logic for appending etcd flags to staticpod to be safer for external etcd
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>.
Feature Gate - Kubeadm Audit Logging
Fixeskubernetes/kubeadm#623
Signed-off-by: Chuck Ha <ha.chuck@gmail.com>
**What this PR does / why we need it**:
This PR enables [Auditing](https://kubernetes.io/docs/tasks/debug-application-cluster/audit/) behind a featureGate. A user can supply their own audit policy with configuration option as well as a place for the audit logs to live. If no policy is supplied a default policy will be provided. The default policy will log all Metadata level policy logs. It is the example provided in the documentation.
**Which issue(s) this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close the issue(s) when PR gets merged)*:
Fixeskubernetes/kubeadm#623
**Special notes for your reviewer**:
**Release note**:
```release-note
kubeadm: Enable auditing behind a feature gate.
```
Audit logs are configurable via the MasterConfiguration file.
All options are ignored unless the FeatureGate is enabled.
Fixeskubernetes/kubeadm#623
Signed-off-by: Chuck Ha <ha.chuck@gmail.com>
Automatic merge from submit-queue (batch tested with PRs 59344, 59595, 59598). 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>.
Fix kubeadm typo
**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
```
With IPv4, the node CIDR prefix is set to /24, which gives 256 pods per node
and 256 nodes, when assuming a /16 is used for the pod subnet.
For IPv6, the node CIDR prefix, is hard coded to /64. This does not work,
because currently the pod subnet prefix must be /66 or higher and must be a
larger subnet (lower value) than the node CIDR prefix.
In addition, the bit mask used to track the subnets (implying the number of
nodes), can only handle 64K entries, so the difference between pod subnet
prefix and node CIDR prefix cannot be more than 16 (bits). The node CIDR
value needs to support this restriction.
To address this, the following algorithm is proposed...
For pod subnet prefixes of /113 or smaller, the remaining bits will be used
for the node CIDR, in multiples of 8, and 9-16 bits will be reserved for the
nodes, so that there are 512-64K nodes and 256, 512, 768, ... pods/node.
For example, with a pod network of /111, there will be 17 bits available. This
would give 8 bits for pods per node and 9 bits for nodes. The node CIDR would
be /120. For a pod network of /104, there will be 24 bits available. There will
be 8 bits for nodes, and 16 bits for pods/node, using a /112 node CIDR.
If the pod subnet prefix is /112, then the node CIDR will be set to /120, and
256 nodes and 256 pods/node will be available.
If the subnet prefix is /113 to /128, we don't have enough bits and will set
the node CIDR prefix to be the same as the pod subnet prefix. This will cause
a falure later, when it tests that the pod subnet prefix is larger than the
node CIDR prefix.
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>.
Enable privileged containers for apiserver and controller
**What this PR does / why we need it**:
In OpenStack environment, when there is no metadata service, we
look at the config drive to figure out the metadata. Since we need
to run commands like blkid, we need to ensure that api server and
kube controller are running in the privileged mode.
**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#47392
Fixes https://github.com/kubernetes/kubeadm/issues/588
**Special notes for your reviewer**:
**Release note**:
```release-note
Fix issue when using OpenStack config drive for node metadata
```
In OpenStack environment, when there is no metadata service, we
look at the config drive to figure out the metadata. Since we need
to run commands like blkid, we need to ensure that api server and
kube controller are running in the privileged mode.
So add a new field in MasterConfiguration for specifying that the
api server and controller manager (s) need extra privileges. Added
a TODO to remove this code when we fully yank out cloud provider
specific calls from these processes.
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>.
Stop using the PersistentVolumeLabel admission controller in v1.9
**What this PR does / why we need it**:
Stop using the PersistentVolumeLabel admission controller in v1.9, as in v1.9, we're targeting out-of-tree cloud providers to be ready to use more generally as part of moving to beta.
**Which issue this PR fixes**:
fixeskubernetes/kubeadm#444
Automatic merge from submit-queue (batch tested with PRs 50832, 51119, 51636, 48921, 51712)
kubeadm: Add support for using an external CA whose key is never stored in the cluster
We allow a kubeadm user to use an external CA by checking to see if ca.key is missing and skipping cert checks and kubeconfig generation if ca.key is missing. We also pass an empty arg --cluster-signing-key-file="" to kube controller manager so that the csr signer doesn't start.
**What this PR does / why we need it**:
This PR allows the kubeadm certs phase and kubeconfig phase to be skipped if the ca.key is missing but all other certs are present.
**Which issue this PR fixes** :
Fixes kubernetes/kubeadm/issues/280
**Special notes for your reviewer**:
@luxas @mikedanese @fabriziopandini
**Release note**:
```release-note
kubeadm: Add support for using an external CA whose key is never stored in the cluster
```
Automatic merge from submit-queue
kubeadm: Add node-cidr-mask-size to pass to kube-controller-manager for IPv6
Due to the increased size of subnets with IPv6, the node-cidr-mask-size needs to be passed to kube-controller-manager. If IPv4 it will be set to 24 as it was previously, if IPv6, it will be set to
64
**What this PR does / why we need it**:
If the user specifies the --pod-network-cidr with kubeadm init, this caused the kube-controller-manager manifest to include the "--allocate-node-cidrs" and "--cluster-cidr" flags to be set. The --node-cidr-mask-size is not set, and currently defaults to 24, which is fine for IPv4, but not appropriate for IPv6. This change passes the a value as the node-cidr-mask-size to the controller-manager. It detects if it is IPv4 or v6, and sets --node-cidr-mask-size to 24 for IPv4 as before, and to 64 for IPv6.
**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes#50469
**Special notes for your reviewer**:
**Release note**:
```release-note
NONE
```
We allow a kubeadm user to use an external CA by checking to see if ca.key is missing and skipping cert checks and kubeconfig generation if ca.key is missing.
Automatic merge from submit-queue (batch tested with PRs 50967, 50505, 50706, 51033, 51028)
kubeadm: Tell the user when a static pod is created
**What this PR does / why we need it**:
Prints a line to notify the user of the static pod creation in order to be consistent with the other phases (one line per phase and optionally per component).
Now the phase command `controlplane all` and `etcd local` also actually outputs something.
Also renamed `[token]` to `[bootstraptoken]` to match the output below and `s/mode/modes/`
`kubeadm init` output now:
```console
$ ./kubeadm init
[kubeadm] WARNING: kubeadm is in beta, please do not use it for production clusters.
[init] Using Kubernetes version: v1.7.4
[init] Using Authorization modes: [Node RBAC]
[preflight] Running pre-flight checks
[preflight] WARNING: docker service is not enabled, please run 'systemctl enable docker.service'
[preflight] Starting the kubelet service
[kubeadm] WARNING: starting in 1.8, tokens expire after 24 hours by default (if you require a non-expiring token use --token-ttl 0)
[certificates] Generated ca certificate and key.
[certificates] Generated apiserver certificate and key.
[certificates] apiserver serving cert is signed for DNS names [thegopher kubernetes kubernetes.default kubernetes.default.svc kubernetes.default.svc.cluster.local] and IPs [10.96.0.1 192.168.1.115]
[certificates] Generated apiserver-kubelet-client certificate and key.
[certificates] Generated sa key and public key.
[certificates] Generated front-proxy-ca certificate and key.
[certificates] Generated front-proxy-client certificate and key.
[certificates] Valid certificates and keys now exist in "/etc/kubernetes/pki"
[kubeconfig] Wrote KubeConfig file to disk: "admin.conf"
[kubeconfig] Wrote KubeConfig file to disk: "kubelet.conf"
[kubeconfig] Wrote KubeConfig file to disk: "controller-manager.conf"
[kubeconfig] Wrote KubeConfig file to disk: "scheduler.conf"
[controlplane] Wrote Static Pod manifest for component kube-apiserver to "/etc/kubernetes/manifests/kube-apiserver.yaml"
[controlplane] Wrote Static Pod manifest for component kube-controller-manager to "/etc/kubernetes/manifests/kube-controller-manager.yaml"
[controlplane] Wrote Static Pod manifest for component kube-scheduler to "/etc/kubernetes/manifests/kube-scheduler.yaml"
[etcd] Wrote Static Pod manifest for a local etcd instance to "/etc/kubernetes/manifests/etcd.yaml"
[init] Waiting for the kubelet to boot up the control plane as Static Pods from directory "/etc/kubernetes/manifests"
[apiclient] All control plane components are healthy after 40.002026 seconds
[markmaster] Master thegopher tainted and labelled with key/value: node-role.kubernetes.io/master=""
[bootstraptoken] Using token: cfe65e.d196614967c3ffe3
[bootstraptoken] Configured RBAC rules to allow Node Bootstrap tokens to post CSRs in order for nodes to get long term certificate credentials
[bootstraptoken] Configured RBAC rules to allow the csrapprover controller automatically approve CSRs from a Node Bootstrap Token
[bootstraptoken] Creating the "cluster-info" ConfigMap in the "kube-public" namespace
[uploadconfig] Storing the configuration used in ConfigMap "kubeadm-config" in the "kube-system" Namespace
[addons] Applied essential addon: kube-dns
[addons] Applied essential addon: kube-proxy
Your Kubernetes master has initialized successfully!
To start using your cluster, you need to run (as a regular user):
mkdir -p $HOME/.kube
sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
sudo chown $(id -u):$(id -g) $HOME/.kube/config
You should now deploy a pod network to the cluster.
Run "kubectl apply -f [podnetwork].yaml" with one of the options listed at:
http://kubernetes.io/docs/admin/addons/
You can now join any number of machines by running the following on each node
as root:
kubeadm join --token cfe65e.d196614967c3ffe3 192.168.1.115:6443 --discovery-token-ca-cert-hash sha256:eb3461b9b707eafc214577f36ae8c351bbc4d595ab928fc84caf1325b69cb192
$ ./kubeadm alpha phase controlplane all
[controlplane] Wrote Static Pod manifest for component kube-apiserver to "/etc/kubernetes/manifests/kube-apiserver.yaml"
[controlplane] Wrote Static Pod manifest for component kube-controller-manager to "/etc/kubernetes/manifests/kube-controller-manager.yaml"
[controlplane] Wrote Static Pod manifest for component kube-scheduler to "/etc/kubernetes/manifests/kube-scheduler.yaml"
$ ./kubeadm alpha phase etcd local
[etcd] Wrote Static Pod manifest for a local etcd instance to "/etc/kubernetes/manifests/etcd.yaml"
```
**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes #
**Special notes for your reviewer**:
**Release note**:
```release-note
NONE
```
@kubernetes/sig-cluster-lifecycle-pr-reviews @fabriziopandini
Due to the increased size of subnets with IPv6, the node-cidr-mask-size needs to be passed to kube-controller-manager. If the user passes a IPv6 cidr, the node-cidr-mask-size will be set to 64, If IPv4 it will be set to 24 as it was previously.
Automatic merge from submit-queue (batch tested with PRs 46458, 50934, 50766, 50970, 47698)
kubeadm: Make the self-hosting with certificates in Secrets mode work again
**What this PR does / why we need it**:
This PR:
- makes the self-hosting with certificates in Secrets mode work
- makes the wait functions timeoutable
- fixes a race condition where the kubelet may be slow to remove the Static Pod
- cleans up some of the self-hosting logic
- makes self-hosting-with-secrets respect the feature flag
**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes #
fixes: https://github.com/kubernetes/kubeadm/issues/405
**Special notes for your reviewer**:
This is work in progress. I'll add unit tests, rebase upon https://github.com/kubernetes/kubernetes/pull/50762 and maybe split out some of the functionatlity here into a separate PR
**Release note**:
```release-note
NONE
```
@kubernetes/sig-cluster-lifecycle-pr-reviews
Automatic merge from submit-queue
kubeadm: Centralize commonly used paths/constants to the constants pkg
**What this PR does / why we need it**:
Before there were constants defined for the control plane components in three different places:
- images
- phases/controlplane
- phases/selfhosting
Now they are in one centralized place. I also moved funcs for building common paths to that lib.
**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes #
Dependency for: https://github.com/kubernetes/kubernetes/pull/48899
**Special notes for your reviewer**:
Most of this PR really is autogenerated with a replace tool. I tested this and things work just normally as well.
**Release note**:
```release-note
NONE
```
@timothysc @dmmcquay @pipejakob @kubernetes/sig-cluster-lifecycle-pr-reviews
Automatic merge from submit-queue (batch tested with PRs 46210, 48607, 46874, 46598, 49240)
kubeadm: Make the hostPath volume mount code more secure
**What this PR does / why we need it**:
- Refactors the hostpath volume mounting code for the Static Pods
- Splits out the functionality that was in a big function to something testable
- Unit test a lot
- Adds support for mounting external etcd CA/cert/key files in an other path than `/etc/ssl/certs`. Before this you **had** to have your files in there or the apiserver would crashloop
- Significantly improves comment coverage
- Now only mounts the bare essentials instead of nearly everything. For example, don't mount full `/etc/kubernetes` when the only thing you need is `/etc/kubernetes/scheduler.conf`
- Make everything but the etcd datadir read-only for components.
**Which issue this PR fixes** *(optional, in `fixes #<issue number>(, fixes #<issue_number>, ...)` format, will close that issue when PR gets merged)*: fixes #
Fixes: https://github.com/kubernetes/kubeadm/issues/341
**Special notes for your reviewer**:
**Release note**:
```release-note
NONE
```
cc @kubernetes/sig-cluster-lifecycle-pr-reviews
