Creating multiple shoot clusters results in EIP being transferred from one cluster to another

[invidian]

How to categorize this issue?

/area networking /kind bug /priority normal /platform equinix-metal

What happened:

When creating 2nd or more shoot cluster, EIPs assigned to first cluster gets re-assigned to new cluster, as currently CCM (v2.0.0) does not tag them uniquely with cluster ID.

What you expected to happen:

After creating 2nd shoot cluster all shoots remain functional.

How to reproduce it (as minimally and precisely as possible):

1. Create 1st shoot, wait for it to converge.
2. Create 2nd shoot, watch the world burn.

Anything else we need to know?:

Recently released CCM v3.0.0 tags EIPs with cluster ID, which should effectively fix this.

Environment:

• Gardener version (if relevant):
• Extension version:
• Kubernetes version (use kubectl version):
• Cloud provider or hardware configuration:
• Others:

[invidian]

I found this 2 issues while trying to update CCM in #113:

• https://github.com/equinix/cloud-provider-equinix-metal/issues/140
• https://github.com/gardener/gardener-resource-manager/issues/107

Since they bring regression to the project (with updated CCM even one Shoots becomes unstable), I think we should wait until we have workarounds for those issues.