Skip to content

Client init fails if one of apis is unavailable #327

@dbazhal

Description

@dbazhal

Hi!
I've faced issue with client init. There are number of custom apis(ApiService resources) in my cluster, like, for example, /apis/metrics.k8s.io/v1beta1. It appeared that not all of apis might work all the time, some apis might stop answering. I'm not sure if that is ok situation, but if such minor api fails, I'd like my cluster's core functionality(running containers, making changes to deployments, etc) continued working.
This discovery code https://github.com/openshift/openshift-restclient-python/blob/master/openshift/dynamic/discovery.py#L108 is pretty good now, but it fails openshift python client in case of one failed minor api(logic "full api discovery failed == client init failed").
Unfortunately I'm not proposing PR to fix that, but I suggest to change that behavior to "some apis discovery failed == continue working and fail if user requested api that is not discovered").

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions