The main advantage is that waiting on channels creates a causal relationship
between goroutines which is visible to synctest. When a controller in a
synctest bubble does a WaitFor in a test's background goroutine for the
controller, the test can use synctest.Wait to wait for completion of cache
sync, without requiring any test specific "has controller synced" API. Without
this, the test had to poll or otherwise wait for the controller.
The polling in WaitForCacheSync moved the virtual clock forward by a random
amount, depending on how often it had to check in wait.Poll. Now tests can be
written such that all events during a test happen at a predictable time. This
will be demonstrated in a separate commit for the
pkg/controller/devicetainteviction unit test.
The benefit for normal production is immediate continuation when the last
informer is synced (not really a problem, but still...) and more important,
nicer logging thanks to the names associated with the thing that is being
waited for. The caller decides whether logging is enabled or disabled and
describes what is being waited for (typically informer caches, but maybe also
event handlers or even something else entirely as long as it implements the
DoneChecker interface).
Before:
Waiting for caches to sync
Caches are synced
After:
Waiting for="cache and event handler sync"
Done waiting for="cache and event handler sync" instance="SharedIndexInformer *v1.Pod"
Done waiting for="cache and event handler sync" instance="SharedIndexInformer *v1.ResourceClaim"
Done waiting for="cache and event handler sync" instance="SharedIndexInformer *v1.ResourceSlice"
Done waiting for="cache and event handler sync" instance="SharedIndexInformer *v1.DeviceClass"
Done waiting for="cache and event handler sync" instance="SharedIndexInformer *v1alpha3.DeviceTaintRule"
Done waiting for="cache and event handler sync" instance="SharedIndexInformer *v1.ResourceClaim + event handler k8s.io/kubernetes/pkg/controller/devicetainteviction.(*Controller).Run"
Done waiting for="cache and event handler sync" instance="SharedIndexInformer *v1.Pod + event handler k8s.io/kubernetes/pkg/controller/devicetainteviction.(*Controller).Run"
Done waiting for="cache and event handler sync" instance="SharedIndexInformer *v1alpha3.DeviceTaintRule + event handler k8s.io/kubernetes/pkg/controller/devicetainteviction.(*Controller).Run"
Done waiting for="cache and event handler sync" instance="SharedIndexInformer *v1.ResourceSlice + event handler k8s.io/kubernetes/pkg/controller/devicetainteviction.(*Controller).Run"
The "SharedIndexInformer *v1.Pod" is also how this appears in metrics.
Kubernetes-commit: fdcbb6cba9a04c028b158bf66d505df7431f63fe
Signed-off-by: Min Jin <minkimzz@amazon.com>
Update staging/src/k8s.io/client-go/tools/cache/the_real_fifo.go
optimizing fifo loop
Co-authored-by: Marek Siarkowicz <marek.siarkowicz@protonmail.com>
Signed-off-by: Min Jin <minkimzz@amazon.com>
refactoring PopBatch to accept []Delta
Signed-off-by: Min Jin <minkimzz@amazon.com>
Kubernetes-commit: 611b4c1408f529de4d4e94e6dd33be2ed1df9276
cache.ErrRequeue advertised itself as a way to requeue failures on a
FIFO, but it suffers the same problems as AddIfNotPresent. If we do
requeue an item at the end, we'll move the informer back in time. If we
requeue at the beginning we'll simply wedge FIFO.
We didn't find examples in the wild, but by removing the error type
those impacted will get a compile error and get to decide what action is
most appropriate for their failure. Most of the time, proceeding to the
next item is best.
Kubernetes-commit: 238c32a1d9b2c72d648193fa8642a53a2884975f
Logically a cache.Queue.AddIfNotPresent means that the informer can move
back in time since an older item is placed after newer items. The
alternative of placing errors at the head of the queue leads to
indefinite memory growth and repeated failures on retry.
Luckily this behavior was behind RetryOnError, which was always set to
false and impossible for normal users to set to true. By removing the
function and setting, impacted users (none found in a github search)
will get a compile failure.
Kubernetes-commit: 8e77ac000131019d5aa49c19aa1f477f6dac4d59
* Add tracker types and tests
* Modify ResourceEventHandler interface's OnAdd member
* Add additional ResourceEventHandlerDetailedFuncs struct
* Fix SharedInformer to let users track HasSynced for their handlers
* Fix in-tree controllers which weren't computing HasSynced correctly
* Deprecate the cache.Pop function
Kubernetes-commit: 8100efc7b3122ad119ee8fa4bbbedef3b90f2e0d
- Run hack/update-codegen.sh
- Run hack/update-generated-device-plugin.sh
- Run hack/update-generated-protobuf.sh
- Run hack/update-generated-runtime.sh
- Run hack/update-generated-swagger-docs.sh
- Run hack/update-openapi-spec.sh
- Run hack/update-gofmt.sh
Signed-off-by: Davanum Srinivas <davanum@gmail.com>
Kubernetes-commit: a9593d634c6a053848413e600dadbf974627515f
Fixes: kubernetes#90581 (the first part)
When `Close()` is invoked on an empty queue, the control loop inside `Pop()` has a small chance of missing the signal and blocks indefinitely due to a race condition. This PR eliminates the race and allows the control loop inside any blocking `Pop()` to successfully exit after Close() is called.
Kubernetes-commit: d8b90955519d10b99415515f8314dd6d35caae8d
2. Remove staging/src/k8s.io/client-go/tools/cache from .golint_failures;
3. Fix some typo from comments.
Kubernetes-commit: 0e0e1f7daba0a6ae6dd59df0a1bb643c323ad8cb
Initial allocation of several maps as the sizes are known
Signed-off-by: Adrián Orive <adrian.orive.oneca@gmail.com>
Kubernetes-commit: 2299f45f52464beff889fce7be1f66642a92b791
