-
Notifications
You must be signed in to change notification settings - Fork 376
"context deadline exceeded" when deleting a pod using openshift #197
Comments
@chavafg any chance you've been able to reproduce it with lower level API ( |
Not really, I see that on kata-runtime logs the call made is And btw, this is also reproducible on kubernetes, using |
Can you list the calls (maybe 10) to |
These are the kata-runtime calls made from the pod creation.
|
Well first thing that strikes me here, the last 2 lines are trying to force delete the same containers. Maybe we are not handling this the way we should but obviously if only one call was made, we would not have any error here. |
hmm, 'double force delete' rings a bell in my head - something we have seen before somewhere that was out of our control - maybe it is a different issue though - @jodh-intel , do you remember if it is maybe relevant, and where we saw that before? |
@grahamwhaley nevermind, I have been talking with @chavafg and the failure happens on the first |
I'll investigate a bit later today on this. I think this might be an issue related to the fact that we |
This commit tries to address uses cases where the caller of the runtime expects a kill to stop the container representing the pod, meaning in our case the end of the VM. This leads to some rework of the delete command since we want a delete --force to always work, even if the container/pod has been killed before. Fixes kata-containers#197 Signed-off-by: Sebastien Boeuf <[email protected]>
This commit tries to address uses cases where the caller of the runtime expects a kill to stop the container representing the pod, meaning in our case the end of the VM. This leads to some rework of the delete command since we want a delete --force to always work, even if the container/pod has been killed before. Fixes kata-containers#197 Signed-off-by: Sebastien Boeuf <[email protected]>
mount: Correct error message with mount failure.
Description of problem
When trying Openshift using CRI-O and Kata-Containers, I can launch workloads, but when trying to delete a pod using:
I see
context deadline exceeded
errors in the logs:Also, after deleting the pod,
kata-runtime list
does not work correctly.Expected result
Pod should be removed correctly
Actual result
Pod is not deleted correctly.
kata-runtime list
does not work correctly.Logs
Openshift node log:
CRI-O Logs:
Kata runtime log:
Kata Proxy Log:
Kata Shim Log:
CRI-O version: 1.9
Openshift version: 3.7.1
Latest Kata containers.
The text was updated successfully, but these errors were encountered: