| Author |
Message
|
| PEPERO |
 Posted: Thu Sep 20, 2012 5:56 am Post subject: Connection Id for the TCPIP connections |
 |
|
Disciple
Joined: 30 May 2011
Posts: 177
|
Hi all;
Is it possible to change the Connection ID of an established TCPIP Server Connection? |
|
| Back to top |
|
 |
| lancelotlinc |
| Posted: Thu Sep 20, 2012 6:21 am Post subject: |
|
|
Jedi Knight
Joined: 22 Mar 2010
Posts: 4941
Location: Bloomington, IL USA
|
No. To get a different connection Id, you must close and reopen. What business requirement is driving the need to change the conx id?
_________________
http://leanpub.com/IIB_Tips_and_Tricks
Save $20: Coupon Code: MQSERIES_READER |
|
| Back to top |
|
|
| NealM |
| Posted: Thu Sep 20, 2012 7:27 am Post subject: |
|
|
Master
Joined: 22 Feb 2011
Posts: 230
Location: NC or Utah (depends)
|
| PEPERO, let me point out that the TCPIPServer's connectionID, LocalEnvironment.TCPIP.Input.ConnectionDetails.Id, serves the same purpose as, and in fact is interchangeable with the MQMD.MsgID, or HTTP or SOAP RequestIdentifier, to uniquely identify to the Broker exactly where to send the reply. (in the old TCPIP support pac, the ID was in fact shorter; When the Nodes were redone and incorporated into the Broker, it was lengthened so as to be interchangeable). So as lancelot said, every time you close a connection the next request will get a new ID. In our high volume asynchronous TCPIPServer flows, we use the connection ID as the MQMD.MsgID/CorrelId on our requests/replies to CICS. And we close the connection "After Data has been Sent" after every xaction (and without any performance issues), so we are getting hundreds of new connection IDs every minute without issue or loss of uniqueness. |
|
| Back to top |
|
|
| PEPERO |
| Posted: Thu Sep 20, 2012 8:33 am Post subject: |
|
|
Disciple
Joined: 30 May 2011
Posts: 177
|
| But as i know we could also determine the msgID when putting the message into the Q (not using MQMI_NONE).So i supposed that the connection ID could also be a user defined string. |
|
| Back to top |
|
|
| lancelotlinc |
| Posted: Thu Sep 20, 2012 8:37 am Post subject: |
|
|
Jedi Knight
Joined: 22 Mar 2010
Posts: 4941
Location: Bloomington, IL USA
|
| PEPERO wrote: |
| But as i know we could also determine the msgID when putting the message into the Q (not using MQMI_NONE).So i supposed that the connection ID could also be a user defined string. |
Why do you want to do this? What business value does it provide you?
_________________
http://leanpub.com/IIB_Tips_and_Tricks
Save $20: Coupon Code: MQSERIES_READER |
|
| Back to top |
|
|
| NealM |
| Posted: Thu Sep 20, 2012 8:59 am Post subject: |
|
|
Master
Joined: 22 Feb 2011
Posts: 230
Location: NC or Utah (depends)
|
| Oi vay! You said TCPIPServerInput Node. This is just like an HTTPInput Node, in that the Broker is in control of what that connection or request ID value is, it sets a unique value, and expects the same value back to complete a request-reply transaction. You wouldn't think of touching the one, why would you think of touching the other??? |
|
| Back to top |
|
|
| PEPERO |
| Posted: Thu Sep 20, 2012 9:06 am Post subject: |
|
|
Disciple
Joined: 30 May 2011
Posts: 177
|
| There is an application using raw TCP/IP sockets interfacing with the SNA gateway to make a connection to host. there is a socket serving stub code running on the SNA gateway which receives the requets with an 8 byte unique number identity code included (in this case it would be the connection ID) for resolving the requests sent to and responses came back from the host. now using the WMB in eliminating the SNA interface we have to map the 8 byte unique number to the connection ID. Since the application creates only a single socket for sending and receiving the requests and responses to the SNA (So there would only be one Connection ID generated by the server connection manager ), i think we have to use the 8 byte identity code instead of the connection id. |
|
| Back to top |
|
|
| lancelotlinc |
| Posted: Thu Sep 20, 2012 9:08 am Post subject: |
|
|
Jedi Knight
Joined: 22 Mar 2010
Posts: 4941
Location: Bloomington, IL USA
|
| PEPERO wrote: |
| we have to map the 8 byte unique number to the connection ID. |
No you don't. These two elements have no relationship to the other.
_________________
http://leanpub.com/IIB_Tips_and_Tricks
Save $20: Coupon Code: MQSERIES_READER |
|
| Back to top |
|
|
| NealM |
| Posted: Thu Sep 20, 2012 12:24 pm Post subject: |
|
|
Master
Joined: 22 Feb 2011
Posts: 230
Location: NC or Utah (depends)
|
| pepero, you should consider those 8 bytes to be part of or all of a unique header that you may have to strip off your data and store someplace (along with the TCPIP connection ID if this is not a synchronous flow) prior to building and sending your CICS request. Upon return from CICS, stick the "header" back on prior to going to the TCPIPServerOutput node. That's rather common processing. But stop thinking you need to do something with the Broker's TCPIP connection ID. It has nothing to do with anything outside of the Broker domain. |
|
| Back to top |
|
|
|
|
