Quick test doesn't seem to run? #624
Labels
Comments
|
Yes this seems to be the same issue as #599 . There is an upstream issue created to try and allow custom taints on nodes, but otherwise each node in your cluster needs to report as schedulable unless it is labeled as a master node. Is that accurate, that you have custom taints/labels on the nodes? You can check with There you can look at all the labels/spec settings. If it doesn't have the label |
|
Our masters have |
|
That does seem OK: see https://github.com/kubernetes/kubernetes/blob/master/test/e2e/framework/util.go#L2750-L2758 The methods for checking the remaining pods check the following:
That is just referring to the last item in the list above. It really just creates a fake pod and checks that it would be tolerated on that node. Effectively a pod without any special tolerations should be able to be scheduled there. |
|
Could you provide the output of |
|
Sure: |
|
@andrewmyhre So the issue is that you have the etcd nodes which are not labled as masters yet have the noSchedule taint on them: So it does seem to be exactly the case #599 is trying to address. To get around this you'd have to either remove the taint or also label those nodes as masters (I'm not positive if there are any other consequences of that). Let me know if I misread something or if my conclusion doesn't seem accurate to you. Otherwise I'll close this in the next day or so. |
johnSchnake
added
misc/duplicate
area/k8s-upstream
and removed
misc/waiting-for-info
labels
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
What steps did you take and what happened:
I'm attempting to run the
--mode=Quicktest(s) but they don't seem to be executing in my cluster. For context, I'm in an air-gapped environment so I'm passing a flag for the sonobuoy image, and both thesonobuoyandsonobuoy-e2e-job-xxxpods do enter a Running state. However they don't appear to do anything.The command I'm executing is:
I then see the pods start:
Viewing the logs for the sonobuoy job pod:
The message
Waiting up to 30m0s for all (but 0) nodes to be schedulableseems relevant.What did you expect to happen:
Not sure.. some logs indicating tests are being run? I'm new to sonobuoy so I'm not sure what to expect.
Anything else you would like to add:
All nodes including workers are healthy and schedulable.
Environment:
v0.13.0v1.12.3/etc/os-release): ubuntu 16.04The text was updated successfully, but these errors were encountered: