October 18, 2011 at 7:59 am
Hi All :-D,
Server 1: Server 2008 SP1, 10.0.2766 (SP1 CU7) - needs updating to , we know.
Server 2: Server 2003 SP2, 9.00.3042.00 (SP2 Vanilla) - needs updating, we know.
We’re seeing problems where service broker messages are not being propagated from the transmission queue on SERVER 1 over to SERVER 2. We are able to kick-start the flow of messages by dropping and creating the Service Broker endpoint on SERVER 1.
-- SERVER 1
USE MASTER;
DROP ENDPOINT ServiceBrokerEndpoint
GO
CREATE ENDPOINT ServiceBrokerEndpoint
AUTHORIZATION [sa]
STATE = STARTED
AS TCP ( LISTENER_PORT = 4022 )
FOR SERVICE_BROKER (
AUTHENTICATION = WINDOWS,
ENCRYPTION = DISABLED);
GO
GRANT CONNECT ON ENDPOINT::ServiceBrokerEndpoint TO PUBLIC;
GO
Further to this, It’s evident that a complete server reboot of SERVER 1 did not prevent this issue reoccurring. That is, we still have an issue where Service Broker for SERVER 1.XXXX appears to be no longer able to send out messages (however it is still able to accept incoming messages). We note that Service Broker sometimes wakes up and makes sporadic outbound connections but a point is reached where even this stops happening. We note also that dropping and recreating the (outbound) endpoint is enough to restart connections but not enough to ensure that they continue.
This is the sys table I've been watching that shows me when outbound TCP connections cease:
-- To prove that the endpoint is working, we should see two or three connections, inbound and outbound (is_accept =0 & =1)
select
t1.principal_name,
t1.last_activity_time,
t1.is_accept, -- 1= connections started remotely
t1.receives_posted, -- network receives that have not yet completed
t1.is_send_flow_controlled, -- 1= network sends have been postponed due to network flow control because the network is busy
t1.is_receive_flow_controlled, -- 1= network receives have been postponed due to flow control because the network is busy
t1.total_bytes_sent,
t1.total_bytes_received,
t1.total_fragments_sent, -- low!
t1.total_fragments_received,
t1.total_sends, -- low on both connections!
t1.total_receives,
t2.* from sys.dm_broker_connections t1
left outer join sys.dm_exec_connections t2 on t1.connection_id = t2.connection_id
Anyone have any thoughts about this one please?
Kind Regards,
Daniel.
October 18, 2011 at 9:13 am
What does message_body in sys.transmission_queue say?
Have you checked that routes are set up correctly on both servers? Make sure that service names (case sensitive), broker id and server names+ports are correct. Can you telnet the remote server + port from both servers? Try running a trace (including all broker events) on both servers to see if there is any communication between them. Any firewall blocking the ports you are using?
October 18, 2011 at 9:14 am
Another question.
Has it ever worked? If so, has anything changed between the time it worked to when it didn't work?
October 18, 2011 at 9:47 am
Hi,
The configuration itself is fine and nothing has changed, its working at the moment but will just fail out of the blue.
Thanks,
Dan
November 10, 2011 at 2:16 pm
Hi Daniel,
I'm having same kind of issue with the below message. Did you come across this message when you were having trouble initiating the message from Server 1 in your example?
"A previously existing connection with the same peer was found after DNS lookup. This connection will be closed. All traffic will be redirected to the previously existing connection. This is an informational message only. No user action is required. State 116."
Thanks,
November 25, 2011 at 11:06 pm
Restarting SQL Server service on the initiator side fixed the issue.
Thanks so much everyone!!
Viewing 6 posts - 1 through 5 (of 5 total)
You must be logged in to reply to this topic. Login to reply