Support celerycam for reconnecting db connenction if db connection is lost accidentally

[jxltom]

Currently if the database is not stable and connection is lost, celerycam can not reconnect db connection automatically and it can not work properly anymore. The error connection already closed will always be raised.

This PR supports to automatically reconnect db connection for celerycam if db connection is lost.

[Copilot]

Pull Request Overview

Adds automatic reconnection logic for Celery snapshot state handling when the Django database connection is lost.

• Imports connection and InterfaceError from Django.
• Wraps worker and task handling in a try/except to catch a closed connection.
• Closes the connection and logs a message so Django’s connection pooling can reconnect.

Comments suppressed due to low confidence (1)

djcelery/snapshot.py:141

• Ensure that a logger instance is defined or imported in this module before using logger.info(), otherwise this will raise a NameError at runtime.

logger.info(

[Copilot]

Comparing exception message strings is brittle; consider catching a more specific exception subclass or inspecting error attributes instead of matching on str(e).

Suggested change

	if str(e) == 'connection already closed':
	if getattr(e, 'args', None) and 'connection already closed' in e.args[0]:

[Copilot]

The try/except block wraps both worker and task handling, which may swallow unrelated errors. Consider narrowing the scope to only the database operations that need reconnection.

Suggested change

	try:
	for worker in state.workers.items():
	self.handle_worker(worker)
	_handle_tasks()
	except InterfaceError as e:
	# When connection already closed exception is raised,
	# force to close connection and Django will automatically reconnect
	if str(e) == 'connection already closed':
	connection.close()
	for worker in state.workers.items():
	self.handle_worker(worker)
	_handle_tasks()
	try:
	# Check and handle database reconnection if needed
	connection.close()
	except InterfaceError as e:
	# When connection already closed exception is raised,
	# force to close connection and Django will automatically reconnect
	if str(e) == 'connection already closed':

[Copilot]

Pull Request Overview

This PR adds automatic reconnection logic for celerycam when the Django database connection is lost.

• Imports connection and InterfaceError to detect closed connections
• Wraps worker/task handling in a try-except to catch InterfaceError and force-close the connection
• Logs when the connection is closed so Django can reconnect on next use

Comments suppressed due to low confidence (1)

djcelery/snapshot.py:136

• Add a unit or integration test that simulates InterfaceError during worker/task handling to verify the reconnection logic runs correctly and doesn’t break the loop.

except InterfaceError as e:

[Copilot]

[nitpick] The try block currently wraps both worker and task loops, which may swallow unrelated InterfaceErrors. Consider narrowing its scope to only the database-close operation or isolating the catch to the specific call that can raise due to a closed connection.

Suggested change

	try:
	for worker in state.workers.items():
	self.handle_worker(worker)
	_handle_tasks()
	except InterfaceError as e:
	# When connection already closed exception is raised,
	# force to close connection and Django will automatically reconnect
	if getattr(e, 'args', None) and 'connection already closed' in e.args[0]:
	connection.close()
	for worker in state.workers.items():
	self.handle_worker(worker)
	_handle_tasks()
	try:
	# Force to close connection if "connection already closed" exception is raised
	connection.close()
	except InterfaceError as e:
	if getattr(e, 'args', None) and 'connection already closed' in e.args[0]:

[Copilot]

Instead of calling connection.close() directly, use django.db.close_old_connections() to properly handle connection pooling and ensure thread safety according to Django's recommendations.

Suggested change

	connection.close()
	close_old_connections()