Don't redeploy worker on Shoot deletion

[timebertt]

How to categorize this PR?

/area quality
/area robustness
/kind bug
/priority normal

What this PR does / why we need it:
We have seen some issues with the DeployWorker step during Shoot deletion, e.g. sometimes the Worker was redeployed in a new run of the deletion flow, although it had already been deleted in a previous run. This sometimes caused resources to leak, that the Worker extension deployed.
The redeployment in the deletion flow was originally introduced to ensure, updated provider credentials were transported to MCM. In the meantime, the generic worker actuator was enhanced to redeploy MachineClasses on Worker deletion and by this also update the provider credentials (ref gardener-attic/gardener-extensions#606).

This PR removes the Worker redeployment in the Shoot deletion flow, because it is not needed anymore and rather might cause issues.

Which issue(s) this PR fixes:
Fixes #2453 (hopefully)

Special notes for your reviewer:

I tested the following scenarios and all worked:

• delete a Shoot with valid credentials
• delete a hibernated Shoot with valid credentials
• delete a Shoot with invalid credentials, update the credentials to valid ones after deletion failure and see it being deleted successfully
• delete a hibernated Shoot with invalid credentials, update the credentials to valid ones after deletion failure and see it being deleted successfully

Release note:

During deletion of a Shoot, the `gardenlet` does not redeploy the `Worker` extension resource anymore, as this sometimes caused leaking resources that blocked the deletion of the Shoot's namespace.

[rfranzke]

/lgtm

Thanks for your extensive tests!