Skip to content

Manila CSI registration failing due to non existing medata service on baremetal nodes #3118

@dgaponcic

Description

@dgaponcic

/kind bug

What happened:
Manila CSI deprecated the --node-id option, and now the value is automatically retrieved from the metadata service.

In our setup, VMs rely on the metadata service, while baremetal nodes rely on config drive. As a result, the CSI fails to register on baremetal nodes:
I0521 08:58:47.584747 1 main.go:111] "Received NotifyRegistrationStatus call" status="error:\"RegisterPlugin error -- plugin registration failed with err: rpc error: code = Internal desc = [NodeGetInfo] unable to retrieve instance id of node error fetching http://169.254.169.254/openstack/latest/meta_data.json: Get \\\"http://169.254.169.254/openstack/latest/meta_data.json\\\": dial tcp 169.254.169.254:80: i/o timeout\""

What you expected to happen:

To either have support for auto discovery from both metadata service and config drive, or keep the --node-id, which we can populate on the k8s side.

How to reproduce it:
Deploy Manila CSI on a node configured with config drive. The registrar container with fail unable to retrieve instance id of node error fetching http://169.254.169.254/openstack/latest/meta_data.json.

Anything else we need to know?:
This also seems to be the case for Cinder CSI.

Environment:

  • openstack-cloud-controller-manager: v1.35.0
  • csi-node-driver-registrar:v2.15.0

Metadata

Metadata

Assignees

No one assigned

    Labels

    kind/bugCategorizes issue or PR as related to a bug.

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions