add example of using CRI proxy
fix: Invalid function call
fix: Optimize getPodImagePullDuration
fix: Return error if the CRI Proxy is undefined
chore: add a document
As a quick fix for a flake, bceec5a3ff
introduced polling with wait.Poll in all callers of CheckDaemonStatus.
This commit reverts all callers to what they were before (CheckDaemonStatus +
ExpectNoError) and implements polling according to E2E best practices
(https://github.com/kubernetes/community/blob/master/contributors/devel/sig-testing/writing-good-e2e-tests.md#polling-and-timeouts):
- no logging while polling
- support for progress reporting while polling
- last but not least, produce an informative failure message in case of a
timeout, including a dump of the daemon set as YAML
The util for checking on daemonstatus was checking once if the Status of
the daemonset was reporting that all the desired Pods are scheduled and
ready.
However, the pattern used in the e2e test for this function was not
taking into consideration that the controller needs to propagate the Pod
status to the DeamonSet status, and was asserting on the condition only
once after waiting for all the Pods to be ready.
In order to avoid more churn code, change the CheckDaemonStatus
signature to the wait.Condition type and use it in a async poll loop on
the tests.
The error was only generated if both checks (generated pods and ready pods)
failed. This looks like a logic error, failing if either of those isn't
matching expectations seems better.
Since we have a Kubernetes-specific fork of go-yaml, use that
consistently throughout the project. This doesn't eliminate the
dependencies on gopkg.in/yaml, which are still used indirectly; but it
ensures that the whole project benefits from fixes or changes to
sigs.k8s.io/yaml.
Signed-off-by: Stephen Kitt <skitt@redhat.com>
It wasn't clear from the comments what "explain" does, leading to calls like
this:
framework.ExpectNoError(fmt.Errorf("additional info ....: %v", ..., err))
* Make JobPodFailurePolicy tests for ignore resilient to random failures
* Increase parallelism and evict in parallel
* Code review fixes to the job e2e tests
The new pull-kubernetes-kind-dra uses
-label-filter='Feature: containsAny DynamicResourceAllocation && !Flaky && !Serial'
to run DRA tests. That didn't work because the E2E framework behind its back
added the default skip expression.
e2e_node test depend on very specific shared state (node state).
Pod leakages between tests oftentimes cause the test preconditions
to be silently corrupted, causing hard to debug CI failures.
We add the option to add an annotation to pods which records
the code line (source code:line) which triggered the pod creation,
so it becomes easier to track which test needs better cleanup.
The relevant e2e framework code is used in all the e2e suites,
so to minimize any unwanted consequences we make the feature
opt-in, planning to enable it initially (and likely only)
in the e2e_node tests.
Signed-off-by: Francesco Romani <fromani@redhat.com>
