[fix][client] Avoid closing reused ServiceUrlProvider in another PulsarClientImpl instance

[oneby-wang]

Motivation

When the same ServiceUrlProvider instance is reused to build a second PulsarClientImpl, the provider rejects the second initialization because each provider instance is tied to one client lifecycle.

The second client constructor currently enters the failure cleanup path and calls shutdown(), which closes conf.getServiceUrlProvider() unconditionally. This can close a provider that is still owned by the first live client.

Modifications

• Track whether the current PulsarClientImpl successfully initialized its ServiceUrlProvider.
• Close the provider during shutdown() only when this client successfully initialized it.
• Add a regression test that verifies a failed second client build does not close the provider used by the first client.

Verifying this change

• Make sure that the change passes the CI checks.

Does this pull request potentially affect one of the following parts:

If the box was checked, please highlight the changes

• Dependencies (add or upgrade a dependency)
• The public API
• The schema
• The default values of configurations
• The threading model
• The binary protocol
• The REST endpoints
• The admin CLI options
• The metrics
• Anything that affects deployment

[void-ptr974]

LGTM. The fix correctly ties ServiceUrlProvider.close() to successful initialization by the current client, and the added regression test covers the reused-provider failure path.