bugfix: stalling test issue (close in TelemetryClientFactory) (#609)

* removed walrus, removed random test before stalling test

Signed-off-by: Sai Shree Pradhan <saishree.pradhan@databricks.com>

* added back random test, connection debug

Signed-off-by: Sai Shree Pradhan <saishree.pradhan@databricks.com>

* formatting

Signed-off-by: Sai Shree Pradhan <saishree.pradhan@databricks.com>

* telemetry client factory debug

Signed-off-by: Sai Shree Pradhan <saishree.pradhan@databricks.com>

* garbage collector

Signed-off-by: Sai Shree Pradhan <saishree.pradhan@databricks.com>

* RLOCK IS THE SOLUTION

Signed-off-by: Sai Shree Pradhan <saishree.pradhan@databricks.com>

* removed debug statements

Signed-off-by: Sai Shree Pradhan <saishree.pradhan@databricks.com>

* remove debugs

Signed-off-by: Sai Shree Pradhan <saishree.pradhan@databricks.com>

* removed debug

Signed-off-by: Sai Shree Pradhan <saishree.pradhan@databricks.com>

* added comment for RLock

Signed-off-by: Sai Shree Pradhan <saishree.pradhan@databricks.com>

---------

Signed-off-by: Sai Shree Pradhan <saishree.pradhan@databricks.com>

Author: saishreeeee

src/databricks/sql/telemetry/telemetry_client.py

@@ -316,7 +316,8 @@ class TelemetryClientFactory:
     ] = {}  # Map of session_id_hex -> BaseTelemetryClient
     _executor: Optional[ThreadPoolExecutor] = None
     _initialized: bool = False
-    _lock = threading.Lock()  # Thread safety for factory operations
+    _lock = threading.RLock()  # Thread safety for factory operations
+    # used RLock instead of Lock to avoid deadlocks when garbage collection is triggered
     _original_excepthook = None
     _excepthook_installed = False