Make the resize test wait until kube-system is healthy #17044
Conversation
|
Labelling this PR as size/M |
k8s-github-robot
added
the
size/M
|
GCE e2e test build/test passed for commit b9e65ce. |
| // Many e2e tests assume that the cluster is fully healthy before they start. Wait until | ||
| // the cluster is restored to health | ||
| By("waiting for system pods to successfully restart") | ||
| pods, err := framework.Client.Pods("kube-system").List(labels.Everything(), fields.Everything()) |
There was a problem hiding this comment.
Wouldn't this lead to some races?
If the new node was just added, kube-system pods that are supposed to be running on that node may not have been created yet - or am I missing something?
There was a problem hiding this comment.
The trouble is that when you kill a node, and add a new one things get rescheduled, so things like influx and kube-dns aren't running yet. And subsequent tests fail because they expect those containers to already be up and running.
|
@wojtek-t friendly ping on this one. thanks! |
|
LGTM |
wojtek-t
added
the
lgtm
|
@k8s-bot test this [submit-queue is verifying that this PR is safe to merge] |
|
GCE e2e build/test failed for commit b9e65ce. |
|
@k8s-bot test this please |
|
GCE e2e test build/test passed for commit b9e65ce. |
|
@k8s-bot unit test this please |
|
@k8s-bot test this [submit-queue is verifying that this PR is safe to merge] |
|
GCE e2e test build/test passed for commit b9e65ce. |
|
Automatic merge from submit-queue |
Auto commit by PR queue bot
Auto commit by PR queue bot
Auto commit by PR queue bot
Peer to #17043