[Bug] Thread leak in NacosRegistry.scheduleServiceNamesLookup due to race condition

### Pre-check

- [x] I am sure that all the content I provide is in English.

### Search before asking

- [x] I had searched in the [issues](https://github.com/apache/dubbo/issues?q=is%3Aissue) and found no similar issues.

### Apache Dubbo Component

Java SDK (apache/dubbo)

### Dubbo Version

Dubbo 3.3.x (Confirmed in latest master branch) & Dubbo 2.7.x.
JDK 1.8+, All Operating Systems.

### Steps to reproduce this issue

1. This issue is identified via static code analysis (concurrency safety check).
2. It is located in `org.apache.dubbo.registry.nacos.NacosRegistry` method `scheduleServiceNamesLookup`.
3. The method uses a "Check-Then-Act" pattern to lazily initialize the `scheduledExecutorService`.
4. Although the field is `volatile`, the initialization block is NOT synchronized.

Code Analysis:
When `scheduleServiceNamesLookup` is called concurrently (e.g. via admin protocol), a race condition occurs:
if (scheduledExecutorService == null) { // Thread A and B can pass here simultaneously
scheduledExecutorService = Executors.newSingleThreadScheduledExecutor(); // Executed twice
// ...
}

### What you expected to happen

The `scheduledExecutorService` should be initialized as a Singleton in a thread-safe manner (e.g., using Double-Checked Locking). Only one thread pool should be created to avoid thread leaks.

### Anything else

### Proposed Solution (Double-Checked Locking)

I suggest modifying the code to use synchronized DCL:

private void scheduleServiceNamesLookup(final URL url, final NacosAggregateListener listener) {
if (scheduledExecutorService == null) {
synchronized (this) {
if (scheduledExecutorService == null) {
// Recommend naming the thread for better debuggability
scheduledExecutorService = Executors.newSingleThreadScheduledExecutor(
new NamedThreadFactory("Dubbo-Nacos-Registry-Scheduler", true)
);
scheduledExecutorService.scheduleAtFixedRate(
() -> {
// ... existing logic ...
},
LOOKUP_INTERVAL,
LOOKUP_INTERVAL,
TimeUnit.SECONDS);
}
}
}
}

### Are you willing to submit a pull request to fix on your own?

- [x] Yes I am willing to submit a pull request on my own!

### Code of Conduct

- [x] I agree to follow this project's [Code of Conduct](https://www.apache.org/foundation/policies/conduct)

Locate org.apache.dubbo.registry.nacos.NacosRegistry and read scheduleServiceNamesLookup, focusing on the lazy scheduledExecutorService initialization and its concurrent callers. Validate the race with the project’s relevant Nacos registry tests or a focused concurrent reproduction; done means concurrent calls create only one scheduler without a thread leak.

Written by the indexing model from the issue text.