-
Notifications
You must be signed in to change notification settings - Fork 127
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
return context timeout error for deallocate #119
base: master
Are you sure you want to change the base?
return context timeout error for deallocate #119
Conversation
Signed-off-by: Periyasamy Palanisamy <[email protected]>
/cc @JanScheurich |
Signed-off-by: Periyasamy Palanisamy <[email protected]>
In https://github.com/containernetworking/cni/blob/spec-v0.4.0/SPEC.md, there is following section and your change violates the CNI specification because your code may return error in DEL action.
In addition, some case, upper CNI plugin expects that only one DEL command is invoked (without error), so the code might not be effective to various CNI plugins. |
Not returning an error means whereabouts accumulates stale IP addresses in its IP pools. It would require an off-line audit function to find and delete such stale IPs. |
@s1061123 thanks for this cni spec pointer! good info. just found an issue with ovs-cni due to missing net ns directory. it needs fixing there. |
Signed-off-by: Periyasamy Palanisamy <[email protected]>
Pull Request Test Coverage Report for Build 1014889463
💛 - Coveralls |
Any chance you could rebase and see if this is still an issue, thanks! |
…m-sync-2023-03-22 Upstream sync 2023 03 29
return context timeout error back to cni/kubelet in case of pod delete, this would allow kubelet to retry for deallocate and avoids stale ip pool entries in the datastore after pod delete.
context timeout errors (due to default 10s request timeout) are frequently seen in highly scaled setup.
Signed-off-by: Periyasamy Palanisamy [email protected]