This change renames NewTypedDelayingQueue to NewTypedDelayingQueue to
stay consistent with the naming scheme in the package. A
NewTypedDelayingQueue constructor is kept for backwards compatibility
but marked as deprecated.
Kubernetes-commit: 45da5dc30ecf0ba4c4b44befd96b8617278c496b
This commit refactors the `renew` function in the leader election package to replace
the deprecated `wait.PollImmediateUntil` with `wait.PollUntilContextTimeout`.
This change simplifies the context handling by removing the need for an additional
timeout context and its cancellation.
The previous implementation created a timeout context for each retry period,
which added unnecessary complexity. The new implementation directly passes the parent
context to the retry function, ensuring that the timeout is handled within `PollUntilContextTimeout`.
This change also ensures that the context passed to `tryAcquireOrRenew` and
`tryCoordinatedRenew` is consistent, improving the readability and maintainability of the code.
Kubernetes-commit: 3d2f498aa31cb2e90bcfc372b7930aed0604fbef
I was workinng on updating a dependency, and noticed that running
hack/update-vendor.sh resulted in a diff. Comitting the result
as a PR.
Signed-off-by: Sebastiaan van Stijn <github@gone.nl>
Kubernetes-commit: aeb607443dd9b8ee378ee10209e9b446256f3ee2
The Error method of the error returned from Request.Watch was "unknown"
even the server returned clear message in the Status struct. It was
because Request.Watch used the Result's err member directly, which is an
unstructured error from the response which the Result object may use if
the caller did not return a structured error.
The patch fixes it by calling the Result's Error method instead, which
returns the structured error when it's present.
It also removes the wrong expectation about events.
Kubernetes-commit: 596c5696c64023808af164284263647d795b0ac2
This makes the Stop method idempotent so that if Stop is called multiple
times, it does not cause a panic due to closing a closed channel.
Signed-off-by: mprahl <mprahl@users.noreply.github.com>
Kubernetes-commit: a54ba917be42c941edf1a0359dced04e1a5e1d6f
When the client does not have permission to watch a resource, the
RetryWatcher continuously retried. In this case, it's better to send an
error and stop retrying to let the caller handle this case since this is
not a transient error that can be recovered without user intervention.
This is particularly helpful in applications that leverage a user
provided service account and the application needs to notify the user to
set the correct permissions for the service account.
This also accounts for invalid credentials from the watch client.
Signed-off-by: mprahl <mprahl@users.noreply.github.com>
Kubernetes-commit: db2218d16e3fb2a30b21ffa0f307beb515d8394b
This is a complete revamp of the original API. Some of the key
differences:
- refocused on structured parameters and allocating devices
- support for constraints across devices
- support for allocating "all" or a fixed amount
of similar devices in a single request
- no class for ResourceClaims, instead individual
device requests are associated with a mandatory
DeviceClass
For the sake of simplicity, optional basic types (ints, strings) where the null
value is the default are represented as values in the API types. This makes Go
code simpler because it doesn't have to check for nil (consumers) and values
can be set directly (producers). The effect is that in protobuf, these fields
always get encoded because `opt` only has an effect for pointers.
The roundtrip test data for v1.29.0 and v1.30.0 changes because of the new
"request" field. This is considered acceptable because the entire `claims`
field in the pod spec is still alpha.
The implementation is complete enough to bring up the apiserver.
Adapting other components follows.
Kubernetes-commit: 91d7882e867da25ae8014f679db32b20e35e89b4
Now all claims are shareable up to the limit imposed by the size of the
"reserverFor" array.
This is one of the agreed simplifications for 1.31.
Kubernetes-commit: 8a629b9f150c1042e2918043e6012a4f22742b19
