* DRA resource claim controller: configurable number of workers
It might never be necessary to change the default, but it is hard to be sure.
It's better to have the option, just in case.
* generate files
* resourceclaimcontroller: normalize validation error message
* Update cmd/kube-controller-manager/app/options/resourceclaimcontroller.go
Co-authored-by: Jordan Liggitt <jordan@liggitt.net>
---------
Co-authored-by: Patrick Ohly <patrick.ohly@intel.com>
Co-authored-by: Jordan Liggitt <jordan@liggitt.net>
kube-controller-manager and kube-scheduler do not use the
configured loopback clients. Drop them.
Fix up scheduler test server to not depend on
the loopback client.
This change refactors the cloud-specific versions of the node lifecycle
and node IPAM controllers to use a context.Context for cancellation and
contextual logging, replacing the legacy stopCh pattern.
This is a follow-up to PR #133985, where these controllers were
separated out due to their use in the legacy Cloud Controller Manager
(CCM).
It is a known issue that the CCM's startup logic does not pass the
controller name via the context. This change proceeds with the
refactoring to unify the cancellation logic across controllers, while
acknowledging that contextual logs will be less detailed when these
controllers are run in the CCM.
Signed-off-by: Aditi Gupta <aditigpta@google.com>
As before when adding v1beta2, DRA drivers built using the
k8s.io/dynamic-resource-allocation helper packages remain compatible with all
Kubernetes release >= 1.32. The helper code picks whatever API version is
enabled from v1beta1/v1beta2/v1.
However, the control plane now depends on v1, so a cluster configuration where
only v1beta1 or v1beta2 are enabled without the v1 won't work.
Add metrics for compatibility version for kube-apiserver, kube-scheduler, and kube-controller-manager.
Signed-off-by: Michael Aspinwall <maspinwall@google.com>
With StreamingCollectionEncodingToJSON and
StreamingCollectionEncodingToProtobuf, the WatchList must re-justify its
necessity. To prevent an ecosystem from building around a feature that
may not be promoted, we will stop serving list-via-watch until
performance numbers can justify its inclusion.
This also stops the kube-controller-manager from using the
list-via-watch by default. The fallback is a regular list, so during
the skew during an upgrade the "right" thing will happen and the new
StreamingCollectionEncoding will be used.
There was one error path that led to a "controller has shut down" log
message. Other errors caused different log entries or are so unlikely (event
handler registration failure!) that they weren't checked at all.
It's clearer to let Run return an error in all cases and then log the
"controller has shut down" error at the call site. This also enables tests to
mark themselves as failed, should that ever happen.
The controller is derived from the node taint eviction controller.
In contrast to that controller it tracks the UID of pods to prevent
deleting the wrong pod when it got replaced.
