StarlingX NFVI Orchestration
ab4b667f9a
If Kubernetes node has images without tag and digest ("crictl images" shows <none>:<none>), taint/untaint node operations in VIM(triggerd by host lock/unlock) that send the requests to Kubernetes to read/patch node via python Kubernetes client will raise this exception. "Invalid value for `names`, must not be `None`" The exception is raised in python Kubernetes client but not from kubernetes. The node is actually patched successfully. Implement a workaround to replace V1ContainerImage.names in the python Kubernetes in the python Kubernete client to bypass the "none image" check. This workaround should be removed if the proposed solutions can be made in kubernetes or a workaround can be implemented in containerd. https://github.com/kubernetes/kubernetes/pull/79018 https://github.com/containerd/containerd/issues/4771 Closes-Bug: 1905481 Signed-off-by: Angie Wang <angie.wang@windriver.com> Change-Id: I747b5309747007204033d8dc2a850d3f824ddc5d |
||
---|---|---|
api-ref/source | ||
devstack | ||
doc | ||
mtce-guest | ||
nfv | ||
nova-api-proxy | ||
releasenotes | ||
.gitignore | ||
.gitreview | ||
.zuul.yaml | ||
centos_build_layer.cfg | ||
centos_dev_docker_images.inc | ||
centos_dev_wheels.inc | ||
centos_guest_image_rt.inc | ||
centos_guest_image.inc | ||
centos_iso_image.inc | ||
centos_pkg_dirs | ||
centos_stable_docker_images.inc | ||
centos_stable_wheels.inc | ||
CONTRIBUTORS.wrs | ||
LICENSE | ||
README.rst | ||
requirements.txt | ||
test-requirements.txt | ||
tox.ini |
nfv
StarlingX NFVI Orchestration