Automatic merge from submit-queue
Add volume reconstruct/cleanup logic in kubelet volume manager
Currently kubelet volume management works on the concept of desired
and actual world of states. The volume manager periodically compares the
two worlds and perform volume mount/unmount and/or attach/detach
operations. When kubelet restarts, the cache of those two worlds are
gone. Although desired world can be recovered through apiserver, actual
world can not be recovered which may cause some volumes cannot be cleaned
up if their information is deleted by apiserver. This change adds the
reconstruction of the actual world by reading the pod directories from
disk. The reconstructed volume information is added to both desired
world and actual world if it cannot be found in either world. The rest
logic would be as same as before, desired world populator may clean up
the volume entry if it is no longer in apiserver, and then volume
manager should invoke unmount to clean it up.
Fixes https://github.com/kubernetes/kubernetes/issues/27653
Automatic merge from submit-queue
CRI: remove pod sandbox resources
The pod-level resources need further discussion. Remove it from CRI for now.
See the original discussion in #29871
Currently kubelet volume management works on the concept of desired
and actual world of states. The volume manager periodically compares the
two worlds and perform volume mount/unmount and/or attach/detach
operations. When kubelet restarts, the cache of those two worlds are
gone. Although desired world can be recovered through apiserver, actual
world can not be recovered which may cause some volumes cannot be cleaned
up if their information is deleted by apiserver. This change adds the
reconstruction of the actual world by reading the pod directories from
disk. The reconstructed volume information is added to both desired
world and actual world if it cannot be found in either world. The rest
logic would be as same as before, desired world populator may clean up
the volume entry if it is no longer in apiserver, and then volume
manager should invoke unmount to clean it up.
Automatic merge from submit-queue
pkg/apiserver/authenticator: reorder oidc plugin to auth after service accounts
Both plugins verify JWTs, but the OpenID Connect plugin performs
much worse when faced with cache misses. Reorder the plugins so
the service account plugin tries to authenticate a bearer token
first.
I had a fun time with this by writing an OpenID Connect provider that stores its data in third party resources. When it's running in the cluster it uses a service account and caused some interesting behavior when the keys expired.
Our OpenID Connect plugin needs a more sophisticated caching model to avoid continuously re-requesting keys when seeing a lot of tokens it doesn't recognize. However, I feel this reordering is generally useful since service accounts will be more common than OpenID Connect tokens.
cc @kubernetes/sig-auth
Automatic merge from submit-queue
Remove pods along with jobs when Replace ConcurrentPolicy is set
Fixes#30442
This builds on #30327 and needs a bit more love in tests.
@janetkuo @erictune fyi
Automatic merge from submit-queue
Add CloudStack cloud provider (extended and refactored)
This PR is superseding PR #26165 is which some groundwork for this PR has been done. So this PR now fixes#26165 and fixes#26045.
I've been in contact with @ngtuna about this updated version of his earlier work (which is still in this PR as one squashed commit) and he has given his 👍 for this 😉
This PR adds additional logic for allocating and associating a public IP, if the `—load-balancer-ip` option is not used. It will do proper management of public IP’s that are allocated by this provider ( so IP’s that are no longer needed/used will also be released again).
Additionally the provider can now also work with CloudStack projects and advanced (VPC) networks. And lastly the Zone interface now returns an actual zone (supplied by the cloud config), a few logical errors are fixed and the first few tests are added.
All the functionality is extensively tested against both basic and advanced (VPC) networks and of course all new and existing (integration) tests are all passing.
Automatic merge from submit-queue
Move new etcd storage (low level storage) into cacher
In an effort for #29888, we are pushing forward this:
What?
- It changes creating etcd storage.Interface impl into creating config
- In creating cacher storage (StorageWithCacher), it passes config created above and new etcd storage inside.
Why?
- We want to expose the information of (etcd) kv client to cacher. Cacher storage uses this information to talk to remote storage.
Automatic merge from submit-queue
move syncNetworkConfig to Init for cni network plugin
start syncNetworkConfig routine in `Init` instead of probing. This fixes a bug where the syncNetworkConfig runs periodically even `cni` network plugin is not in use.
Automatic merge from submit-queue
add RequiresExactMatch test for empty andterm
What?
Add a test path for empty andterm.
Why?
fields.Everything() returns empty andterm.
fields.SelectorFromSet() returns empty andterm.
Automatic merge from submit-queue
Remove resource specifications from CRI until further notice
See #29871 for the discussion issue.
cc @dchen1107 @vishh @yujuhong @euank @yifan-gu @feiskyer
Automatic merge from submit-queue
Endpoint controller logs errors during normal cluster behavior
The endpoint controller logs an error when its forbidden from creating new endpoints during namespace termination. This is normal cluster behavior, and therefore should not be logged. This confuses operators administrating the cluster.
Updated to log at a lower level in response to a forbidden message when performing a create operation. In case of an error on the API server side of the house, I continue to requeue the key. It should be ignored in a future syncService call once the service is deleted as part of namespace termination.
See https://bugzilla.redhat.com/show_bug.cgi?id=1347425
/cc @kubernetes/rh-cluster-infra
This commit adds logic for allocating and associating a public IP, if the `—load-balancer-ip` option is not used. It will do proper management of IP’s that are allocated by this provider, so IP’s that are no longer needed/used will also be released again.
Additionally the provider can now also work with CloudStack projects and advanced (VPC) networks.
Lastly the Zone interface now returns an actual zone (supplied by the cloud config), a few logical errors are fixed and the first few tests are added.
All the functionality is extensively tested against both basic and advanced (VPC) networks.
Automatic merge from submit-queue
speed up RC scaler
The RC scaler was waiting before starting the scale and then didn't use a watch to observe the result. That led to longer than expected wait times.
@fabianofranz ptal. You may want to sweep the rest of the file. It could use some tidying with `RetryOnConflict` and `watch.Until`.
Automatic merge from submit-queue
Validate SHA/Tag when checking docker images
Docker API does not validate the tag/sha, for example, all the following
calls work say for a alpine image with short SHA "4e38e38c8ce0"
echo -e "GET /images/alpine:4e38e38c8ce0/json HTTP/1.0\r\n" | nc -U /var/run/docker.sock
echo -e "GET /images/alpine:4e38e38c/json HTTP/1.0\r\n" | nc -U /var/run/docker.sock
echo -e "GET /images/alpine:4/json HTTP/1.0\r\n" | nc -U /var/run/docker.sock
So we should check the response from the Docker API and look for the tags or SHA explicitly.
Fixes#30355
Automatic merge from submit-queue
Make more messages respect --quiet flag
Make following two messages respect `--quiet` in `kubectl run`
- `If you don't see a command prompt, try pressing enter.`
- `Pod "name" deleted`
Ref #28695
Automatic merge from submit-queue
Set pod state as "unknown" when CNI plugin fails
Before this change, CNI plugin failure didn't change anything in the pod status, so pods having containers without requested network were "running".
Fixes#29148
Automatic merge from submit-queue
Remove kubelet pkill dependency
Issue #26093 identified pkill as one of the dependencies of kublet
which could be worked around. Build on the code introduced for pidof
and regexp for the process(es) we need to send a signal to.
Related to #26093
Automatic merge from submit-queue
Adding events to federation control plane
Adding events to federation control plane.
Apart from the standard changes to add a resource to `federation/apis/core/v1`, other changes are:
* Adding a new `federationoptions.ServerRunOptions` which includes `genericoptions.ServerRunOptions` and EventsTTL.
* Added a new method in `pkg/api/mapper` to build a RestMapper based on the passed Scheme rather than using `api.Scheme`. Updated `federation/apis/core/install` to use this new method. Without this change, if `federation/apis/core/install.init()` is called before `pkg/api/install.init()` then the registered RESTMapper in `pkg/apimachinery/registered` will have no resources. This second problem will be fixed once we have instances of `pkg/apimachinery/registered` instead of a single global singleton (generated clientset which imports `pkg/api/install` will have a different instance of registered, than federation-apiserver which imports `federation/apis/core/install`).
cc @kubernetes/sig-cluster-federation @lavalamp
Automatic merge from submit-queue
[kubelet] Introduce --protect-kernel-defaults flag to make the tunable behaviour configurable
Let's make the default behaviour of kernel tuning configurable. The default behaviour is kept modify as has been so far.
Automatic merge from submit-queue
Fix TestPidOf {procfs} - Take #2
We should not bailout when we get an error. We should continue
processing other files/directories. We were returning the
err passed in which was causing the processing to stop.
Fixes#30377
Automatic merge from submit-queue
Kubelet: generate sandbox/container config for new runtime API
Generate sandbox/container config for new runtime API. Part of #28789 .
CC @yujuhong @Random-Liu @dchen1107
Automatic merge from submit-queue
Add zsh compatibility note `completion` cmd help
zsh completions are not supported on zsh versions < 5.2.
This patch advices user on supported versions of zsh when using the `completion`
command to avoid potential UX failure.
##### After
`$ kubectl completion -h`
```
Output shell completion code for the given shell (bash or zsh).
This command prints shell code which must be evaluation to provide interactive
completion of kubectl commands.
Examples:
$ source <(kubectl completion bash)
will load the kubectl completion code for bash. Note that this depends on the
bash-completion framework. It must be sourced before sourcing the kubectl
completion, e.g. on the Mac:
$ brew install bash-completion
$ source $(brew --prefix)/etc/bash_completion
$ source <(kubectl completion bash)
If you use zsh*, the following will load kubectl zsh completion:
$ source <(kubectl completion zsh)
* zsh completions are only supported in versions of zsh >= 5.2
```
```release-note
release-note-none
```
<!-- Reviewable:start -->
---
This change is [<img src="https://reviewable.kubernetes.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.kubernetes.io/reviews/kubernetes/kubernetes/30460)
<!-- Reviewable:end -->
Kris
