Checking TimeoutError from get_connection in _execute_cluster_command

[EugeneChung]

• _execute_command
  Both ConnectionError and TimeoutError are catched and it reinitializes the cluster topology.
  

  valkey-py/valkey/cluster.py

  Lines 1149 to 1180 in a4d180b

  	connection = get_connection(valkey_node, *args, **kwargs)
  	if asking:
  	connection.send_command("ASKING")
  	valkey_node.parse_response(connection, "ASKING", **kwargs)
  	asking = False
  	response_from_cache = connection._get_from_local_cache(args)
  	if response_from_cache is not None:
  	return response_from_cache
  	else:
  	connection.send_command(*args)
  	response = valkey_node.parse_response(connection, command, **kwargs)
  	if command in self.cluster_response_callbacks:
  	response = self.cluster_response_callbacks[command](
  	response, **kwargs
  	)
  	if keys:
  	connection._add_to_local_cache(args, response, keys)
  	return response
  	except AuthenticationError:
  	raise
  	except (ConnectionError, TimeoutError) as e:
  	# Connection retries are being handled in the node's
  	# Retry object.
  	# ConnectionError can also be raised if we couldn't get a
  	# connection from the pool before timing out, so check that
  	# this is an actual connection before attempting to disconnect.
  	if connection is not None:
  	connection.disconnect()
  	# Remove the failed node from the startup nodes before we try
  	# to reinitialize the cluster
  	self.nodes_manager.startup_nodes.pop(target_node.name, None)

• _send_cluster_commands
  It only catchs ConnectionError. Retry with the cluster topology update doesn't work for TimeoutError.
  

  valkey-py/valkey/cluster.py

  Lines 2106 to 2121 in a4d180b

  	retry_attempts = self.cluster_error_retry_attempts
  	while True:
  	try:
  	return self._send_cluster_commands(
  	stack,
  	raise_on_error=raise_on_error,
  	allow_redirections=allow_redirections,
  	)
  	except (ClusterDownError, ConnectionError) as e:
  	if retry_attempts > 0:
  	# Try again with the new cluster setup. All other errors
  	# should be raised.
  	retry_attempts -= 1
  	pass
  	else:
  	raise e

  
  

  valkey-py/valkey/cluster.py

  Lines 2177 to 2181 in a4d180b

  	try:
  	connection = get_connection(valkey_node, c.args)
  	except ConnectionError:
  	for n in nodes.values():
  	n.connection_pool.release(n.connection)

I think ClusterPipeline.ERRORS_ALLOW_RETRY should be considered.

[mkmkme]

Hey! Thanks for the report! Can you please check if #284 fixes the issue for you? Thanks!

[EugeneChung]

@mkmkme I'm going to test it and let you know.

[EugeneChung]

@mkmkme I think it's correctly fixed. Thanks! Please let me know if the fix release comes out.

[mkmkme]

@EugeneChung it will be released with 7.0.0. I hope to start releasing beta builds for it quite soon, hopefully within a month.