Open tylerfanelli opened 5 months ago
@wainersm You mentioned that you have ran some of these tests on Fedora before. Have you encountered this issue?
@tylerfanelli sometime back @c3d suggested the following to disable swap on Fedora, otherwise it comes back.
dnf remove zram-generator-defaults
@bpradipt Thanks for pointing this out, your fix seems to have solved my problem.
I see that the image was built:
$ sudo docker images
REPOSITORY TAG IMAGE ID CREATED SIZE
localhost:5000/cc-operator latest 48f322b96469 35 minutes ago 54.2MB
Yet I don't think it is ran:
$ sudo docker ps
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
ab56de994a83 registry:2.8.1 "/entrypoint.sh /etc…" 39 minutes ago Up 39 minutes 0.0.0.0:5000->5000/tcp local-registry
What is the difference between cc-operator
and local-registry
? Is this expected? As in, do I still have to run the operator with docker run ... localhost:5000/cc-operator ...
, or should it already be running after the script exits?
The cc-operator is the newly built operator image which is pushed to the local docker registry. That's the one you see with container id ab56de994a83
.
The tests create cc-operator pod on the K8s cluster using the cc-operator image in the local registry. You don't need to run it manually as run-local.sh takes care of it.
Describe the bug On Fedora, when running
tests/e2e/run-local.sh
, when bringing up the local k8s cluster, an error occurs that prevents the cluster from being brought up.To Reproduce Steps to reproduce the behavior:
Describe the results you received: All previous checks pass, yet on
INFO: Bring up the test cluster
, I'm met with the following:Notice the warning:
[WARNING Swap]: swap is enabled; production deployments should disable swap unless testing the NodeSwap feature gate of the kubelet
I've looked up the error message
[kubelet-check] It seems like the kubelet isn't running or healthy.
and found this post discussing the issue. One popular answer indicates that the issue is due to swap being enabled, which from the previous warning I mentioned, seems like it could be a likely cause of my problem.So, I undo the changes made from the test:
$ ./run-test.sh -u
Check the current swaps:
Disable swap:
Check to ensure the swaps are gone:
Seems that they are gone, so I run
run.sh
again:$ ./run.sh -r kata-qemu-snp
And am met with the following:
Seeing that the system reports swap being enabled again, I view the swaps being used:
Is there something in the system (k8s, systemd?) that is re-initializing swaps? I ensure they are disabled before the test is ran; yet after, they are re-enabled.
Has anyone seen this before or have any potential solutions? Thanks.