Otherwise, nodeNameToPodList[nodeName] list will have all its references
identical (corresponding to the control variable reference).
Thus, making all the pods in the list identical.
The purpose of the pod created by `createBalancedPodForNodes()` is to ensure
that all nodes have equal resource requests (as seen by the scheduler). This
prevents the default scheduling behavior (which attempts to balance resource requests)
from interfering with e2e's which test other priorities/score plugins.
Because the scheduler only worries about requests, specifying `Limits` in this pod
is unnecessary. In fact, if the calculated "balancing" limit is too low, it can cause
the balancing pod to never start due to OOMKill errors, leading to flakes and failures.
The test is not cleaning all pods it created.
Memory balancing pods are deleted once the test namespace is.
Thus, leaving the pods running or in terminating state when a new test is run.
In case the next test is "[sig-scheduling] SchedulerPredicates [Serial] validates resource limits of pods that are allowed to run",
the test can fail.
WaitForStableCluster() checks all pods run on worker nodes, and the
function used to refer master nodes to skip checking controller plane
pods.
GetMasterAndWorkerNodes() was used for getting master nodes, but the
implementation is not good because it usesDeprecatedMightBeMasterNode().
This makes WaitForStableCluster() refer worker nodes directly to avoid
using GetMasterAndWorkerNodes().
WaitForPod*() are just wrapper functions for e2epod package, and they
made an invalid dependency to sub e2e framework from the core framework.
So this replaces WaitForPodRunning() with the e2epod function.
This PR moves functions from test/e2e/framework.util.go for making e2e
core framework small and simple:
- RestartKubeProxy: Moved to e2e network package
- CheckConnectivityToHost: Moved to e2e network package
- RemoveAvoidPodsOffNode: Move to e2e scheduling package
- AddOrUpdateAvoidPodOnNode: Move to e2e scheduling package
- UpdateDaemonSetWithRetries: Move to e2e apps package
- CheckForControllerManagerHealthy: Moved to e2e storage package
- ParseKVLines: Removed because of e9345ae5f0
- AddOrUpdateLabelOnNodeAndReturnOldValue: Removed because of ff7b07c43c
Includes the changes from #80922 (that were reverted), and updates the test to appropriately add intolerable taints to all nodes except the target node
I think, if a pod doesn't have any tolerations, we don't prefer node without taints to
the one which has taints in https://github.com/kubernetes/kubernetes/blob/master/pkg/scheduler/algorithm/priorities/taint_toleration.go#L29, so there is no point in testing that particular functionality. The side effect of the above is, since we're going round-robin in every scheduling cycle sometimes we're choosing first node and in the next cycle we'd move onto next node(where taints are not being applied), so it's causing problem unnecessarily
- Add a package "node" under e2e/framework and alias e2enode;
- Rename some functions whose name have redundant string.
Signed-off-by: Jiatong Wang <wangjiatong@vmware.com>
