[cloud-provider-vcd] Add legacy mode support for VCD cloud provider CCM

[ptrvsrg]

Description

Fixes the CCM controller list rendering for cloud-provider-vcd. The --controllers flag was built with a broken printf format string ("%s,$s" instead of "%s,%s"), producing an invalid value like --controllers=cloud-node,$s%!(EXTRA string=cloud-node-lifecycle).

Also restores support for legacy CCM controller names for VCD Legacy mode:

• legacy (legacyMode: true): cloud-node, cloud-node-lifecycle, and service (when the load balancer is enabled);
• current: cloud-node-controller, cloud-node-lifecycle-controller, and service-lb-controller.

Why do we need it, and what problem does it solve?

On clusters installed from the main branch, the VCD CCM was in CrashLoop because the --controllers argument was malformed and old CCM controller names were no longer supported for VCD Legacy. This change fixes the malformed argument and returns the legacy controller names, so CCM works for
both legacy and current VCD API versions.

Why do we need it in the patch release (if we do)?

Not necessarily

Checklist

• The code is covered by unit tests.
• e2e tests passed.
• Documentation updated according to the changes.
• Changes were tested in the Kubernetes cluster manually.

Changelog entries

section: cloud-provider-vcd
type: fix
summary: Fix malformed CCM `--controllers` argument and restore legacy controller names for VCD Legacy, resolving cloud-controller-manager CrashLoop.
impact_level: default

[deckhouse-BOaTswain]

⚪ e2e: VCD for deckhouse:fix/vcd-legacy-ccm-controllers cancelled.

Workflow details

[deckhouse-BOaTswain]

🟢 e2e: VCD for deckhouse:fix/vcd-legacy-ccm-controllers succeeded in 34m38s.

Workflow details

VCD-Standard-Containerd-1.34 - Connection string: ssh -J ubuntu@80.249.129.56 ubuntu@192.168.254.216

🟢 e2e: VCD, Containerd, Kubernetes 1.34 succeeded in 33m27s.