yahoo-eng-team team mailing list archive
-
yahoo-eng-team team
-
Mailing list archive
-
Message #84703
[Bug 1906759] [NEW] Port remains reserved even after deleting the instance managed via kubernetes
Public bug reported:
Delete instance with a Neutron port. Create an instance with the same port. Port will remain in the wrong state ("in use" with deleted instance UUID).
The creation of the new instance will fail. Port should be released, not stay.
Here, I'm managing the OpenStack instances directly using Kubernetes.
Also this a random problem and not happens every time.
Environment details:
k8s: kops (https://github.com/kubernetes/kops) to deploy kubernetes
CNI: Calico
Kubernetes version 1.19.3 specifically
A fix has been already done to kops: https://github.com/kubernetes/kops/pull/10178
But there is still a bug with OpenStack API's which is causing race
condition and the port is getting left to reserved state even after the
instance is deleted.
Considering this as a production environment, there was a little chance
to reproduce and capture the logs.
** Affects: nova
Importance: Undecided
Status: New
--
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (nova).
https://bugs.launchpad.net/bugs/1906759
Title:
Port remains reserved even after deleting the instance managed via
kubernetes
Status in OpenStack Compute (nova):
New
Bug description:
Delete instance with a Neutron port. Create an instance with the same port. Port will remain in the wrong state ("in use" with deleted instance UUID).
The creation of the new instance will fail. Port should be released, not stay.
Here, I'm managing the OpenStack instances directly using Kubernetes.
Also this a random problem and not happens every time.
Environment details:
k8s: kops (https://github.com/kubernetes/kops) to deploy kubernetes
CNI: Calico
Kubernetes version 1.19.3 specifically
A fix has been already done to kops: https://github.com/kubernetes/kops/pull/10178
But there is still a bug with OpenStack API's which is causing race
condition and the port is getting left to reserved state even after
the instance is deleted.
Considering this as a production environment, there was a little
chance to reproduce and capture the logs.
To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1906759/+subscriptions
Follow ups