The user expectections calling this method is that the pod should
be ready for the test, however, it only checks that is running,
causing timing issues on busy environments.
Per example, if the pod is not ready, kube-proxy or other services
implementations will not forward traffic to it.
A number of e2e tests are useful to run after the system has been
disrupted or is in the progress of being disrupted, but the current
suite and test logic blocks progress waiting for all nodes to be
healthy.
By passing -1 to --minStartupPods or --allowed-not-ready-nodes flags
the caller can bypass wait logic before and after test suites that
would prevent running e2e during disruption. This allows use of parts
of the e2e suite during cluster duress to verify that controllers or
components still function.
WaitForPodSuccessInNamespace[Slow] are replaced by WaitForPodSuccessInNamespaceTimeout(),
so that custom timeouts are used instead of the hardcoded ones.
On HA API server hiccups we saw a prow job with error:
fail [k8s.io/kubernetes@v1.19.0/test/e2e/e2e.go:284]: Sep 30 17:06:08.313: Error waiting for all pods to be running and ready: 0 / 0 pods in namespace "kube-system" are NOT in RUNNING and READY state in 10m0s
POD NODE PHASE GRACE CONDITIONS
The failure should include the last error from API server, if it's
available:
fail [k8s.io/kubernetes@v1.19.0/test/e2e/e2e.go:284]: Oct 1 11:29:45.220: Error waiting for all pods to be running and ready: 0 / 0 pods in namespace "kube-system" are NOT in RUNNING and READY state in 10m0s
Last error: Get "https://localhost:6443/api/v1/namespaces/kube-system/replicationcontrollers": dial tcp [::1]:6443: connect: connection refused
POD NODE PHASE GRACE CONDITIONS
Some of the tests are negative test cases which are supposed to ensure that those
invalid usecases are handled properly.
However, some of the tests are false positives, they can pass due to various reasons.
One such example is: "should fail substituting values in a volume subpath with absolute path".
This test can pass if:
- the Pod cannot start due to various reasons (e.g.: the container image cannot be pulled or does
not exist).
- the Pod ran to completion, even though the container was not supposed to start in the first place.
Organized functions that abstract the access of
k8s.io/apimachinery/pkg/runtime.Objects into a framework subpackge.
Signed-off-by: Jorge Alarcon Ochoa <alarcj137@gmail.com>
Because of a := assignment, the anonymous function assigned the pod
list to a local variable instead of the
WaitForPodsWithLabelRunningReady return value which therefore always
was nil.
The correct code is an assignment with = as in WaitForPodsWithLabelScheduled.
