| Author |
Message
|
| NMurn_5 |
 Posted: Thu Mar 01, 2007 8:25 am Post subject: AMQ9208: and TCP/IP Return Code 73 |
 |
|
Newbie
Joined: 11 Aug 2006
Posts: 7
|
Hello everyone!
I searched for a while on both of the errors I'm receiving and couldn't find much. First let me explain the systems I'm working with:
I'm working with WebSphere Application Server 6.1.0.2 running on an AIX box which is version 5.3 ML5. We have MQ Client 6.0.1.1 on this box which connects to the qmgr remotely on another box which has MQ Server 6.0.1.1 on it. And finally we have a zOS mainframe in the background.
Here's what's happening:
We have a web application that upon logging into, sends a string to MQ which sends to the mainframe. The data sent back populates the two drop down menus on the web application. Currently these two drop downs are not populating. When I checked the application log I see a "java.lang.NullPointerException" right after the initial message is sent to MQ for both of the drop downs.
The MQ logs show the following :
03/01/07 10:00:35 - Process(155706.308) User(root) Program(amqrmppa)
AMQ9208: Error on receive from host pcxxxxxx (xx.x.xx.xxx).
EXPLANATION:
An error occurred receiving data from pcxxxxxx (xx.x.xx.xxx) over TCP/IP. This
may be due to a communications failure.
ACTION:
The return code from the TCP/IP (read) call was 73 (X'49'). Record these values
and tell the systems administrator.
I looked in the errno.h file and saw the return code 73 had the following:
#define ECONNRESET 73 /* Connection reset by peer */
Which I found out happens when a client connection is terminated without cleanly disconnecting.
I can take the message being sent by the application and manually put in our remote queue and I can see the local response queue get the messages. I browsed them and they were all there and correct.
After this wall o' text(sorry), my main question, is does this seem like it's a code problem with the application or does it seem like a network error? I've pretty much ruled out network error because there are 3+ other applications on this box that are working fine with MQ as I type this.
I feel as if though I'm missing something super small and I'll probably scream when I find it. If anyone needs any additional information please don't hesitate to ask.
Any help/insight is very much appreciated. |
|
| Back to top |
|
 |
| jefflowrey |
| Posted: Thu Mar 01, 2007 8:32 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
First rule of MQ is ... elminate application problems first... 
Likely the null pointer exceptions are causing the connection drops, as the code is probably not written to finalize the handles when it needs to.
It might not be the Java program though, it could be that the mainframe isn't sending back the right responses with the right ids within the right timeframe.
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| NMurn_5 |
| Posted: Thu Mar 01, 2007 8:44 am Post subject: |
|
|
Newbie
Joined: 11 Aug 2006
Posts: 7
|
| jefflowrey wrote: |
It might not be the Java program though, it could be that the mainframe isn't sending back the right responses with the right ids within the right timeframe. |
Would you mind explaining what you wrote about the mainframe ids a little more? I'm not a mainframe person 
I have one of the mainframe women sitting about 2 chairs away from me so I could have her check on her side, I just don't know what it is I should ask her lol.
When you say ids are you talking about the sequence numbers or something completely different? I had her check her channels on her side and she said there was nothing wrong with them.
Thanks for the quick response |
|
| Back to top |
|
|
| jefflowrey |
| Posted: Thu Mar 01, 2007 9:03 am Post subject: |
|
|
Grand Poobah
Joined: 16 Oct 2002
Posts: 19981
|
I mean mainly Correlation ID and Mesage ID on the MQMD
_________________
I am *not* the model of the modern major general. |
|
| Back to top |
|
|
| fjb_saper |
| Posted: Fri Mar 02, 2007 8:38 pm Post subject: |
|
|
Grand High Poobah
Joined: 18 Nov 2003
Posts: 20756
Location: LI,NY
|
Have you checked about the app server (WAS) handling correctly the resources and their transactionality as well as their release (back to the pool)? Could it be that you are running out of resources in the pool ?
Remember that WAS overwrites any transactionality settings on the MQSession....
Enjoy 
_________________
MQ & Broker admin |
|
| Back to top |
|
|
| NMurn_5 |
| Posted: Mon Mar 05, 2007 6:04 am Post subject: |
|
|
Newbie
Joined: 11 Aug 2006
Posts: 7
|
| fjb_saper wrote: |
Have you checked about the app server (WAS) handling correctly the resources and their transactionality as well as their release (back to the pool)? Could it be that you are running out of resources in the pool ?
Remember that WAS overwrites any transactionality settings on the MQSession....
Enjoy |
Hey fjb,
We've gone over the WAS settings 100 times. The thing is, we just recently upgraded WAS from 5.1.1.4 to 6.1.0.2 and we just upgraded MQ to version 6.0.1.1. We basically mirrored everything from the old systems(settings wise) to the new systems. My first thought was there was an issue with some of the settings, but there are 5 other applications on the same JVM/App Server that are communicating with MQ without a problem and this one application is the only one having an issue.
After going over the logs and checking MQ it leads me to believe it's how the developers are having the application connect to MQ because when they initially login I see no connection in MQ happening in the logs and the INQ Queue never connects with the mainframe. I'm actually going to have a talk with them today and go over their parameters as far as their initial connection goes. I'll post back what I find.
Thanks. |
|
| Back to top |
|
|
|
|
