redis: preserve the source endpoint in AsyncDBUpdater

[ashutosh-agrawal]

What I changed

• Changed AsyncDBUpdater to create its worker connection with m_db->newConnector(0).
• The worker still gets its own Redis connection, but it now uses the supplied connector's actual endpoint: either its Unix socket path or its TCP host and port.
• Added a focused test that starts with a Unix socket connection, queues an asynchronous update, and verifies that the value reaches Redis.

Why

AsyncDBUpdater performs background database writes, including writes used by ZMQ state-table persistence.

The caller can connect to Redis through a Unix socket. However, AsyncDBUpdater previously rebuilt its worker connection with the TCP option forced on. This created the following failure:

1. The caller's initial Unix socket connection succeeds.
2. AsyncDBUpdater creates a separate TCP connection for queued writes.
3. The service therefore still depends on the Redis TCP listener.
4. If the TCP listener is disabled, asynchronous database updates fail even though the caller connected successfully.

Creating the worker with newConnector() preserves the endpoint already selected by the caller and removes that mismatch.

How the test covers the fix

• Confirms that the supplied connector uses a Unix socket.
• Queues an update through AsyncDBUpdater and waits for the worker to process it.
• Reads the value back through the supplied connector and verifies the written field and value.

[azure-pipelines]

Azure Pipelines:
There may be pipelines that require an authorized user to comment /azp run to run.

[mssonicbld]

/azp run

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).

[mssonicbld]

/azp run

[azure-pipelines]

Azure Pipelines:
Successfully started running 1 pipeline(s).