Question

We are consuming TM1 cubes with Report Studio through Framework Manager. Quite often when I am trying to come up with new solutions to my challenges in Report Studio, I get an error when I run the report, and then the server goes down. Then I have to restart the dispatchers (Cognos Administration -> Status -> System -> Right Click on the server -> Test Dispatchers -> Right Click on the server -> Start Dispatchers). The error message that I get is:

The connection closed before the request is processed. If you are using WebSphere Application Server, to reduce the frequency of this error, increase the Persistent Timeout parameter for the Web container transport chains in the administrative console. Increase the time in 10-15 second intervals until the error no longer or rarely occurs.

We are not using WebSphere, but Tomcat (default with the installation).
-> Increasing connection timout interval on WebSphere thus not applicable
-> The timeout interval in the Tomcat config seems to be 60 seconds (60000 ms)

More importantly: The error message shows immediately (after 1 second) when I run the report
-> Indicates to me that this is regardless of any timeout interval setting

Additional info: The error message comes almost always when I manually and dynamically attempt to build MUNs. However, sometimes (dunno when and why) it shows the MUN that I've created and tells me that it is invalid. Which is WAY much better for debugging.

Any suggestions on why this is happening and how to fix it would be greatly appreciated!


Edit 1: http://www.linkedin.com/groups/Product-Cognos-BI-1011-Cognos-3917273.S.143157206 This post states (almost at the bottom) that

When the Cognos BI report ask for a field that does not exist, the TM1 Application disconnects the connection. And the Cognos BI Report will timeout.

Is this true? If so; why am I sometimes told that my MUN is invalid, whereas other times the connection is closed and the server shut down? Is it because even Report Studio thinks that my MUN is valid and tries to get it from the TM1 Server?

And additionally: Is it possible to change this behavior for the TM1 server?

Edit 2: Or change the BI server behavior so that it does not shut down when the TM1 connection is disconnected, but rather show an error of some kind? Thanks again!

Edit 3: Okay, so I did some checking with the TM1 top utility (http://pic.dhe.ibm.com/infocenter/ctm1/v9r5m0/index.jsp?topic=%2Fcom.ibm.swg.im.cognos.tm1_op.9.5.1.doc%2Ftm1_op_id6961UsingtheTM1TopUtility_N160F47.html). When a normal report is run, a new thread is shown in the monitoring list. This thread then disappears when I stop the BI server dispatchers, or automatically after approximately 5 minutes of idle time without any reports being run (according to the TM1 Top log dump).

Likewise, when the error occurs, a new thread is shown in the list. However, it disappears after a short second (probably because the BI server dispatchers are shut down).

I have therefore concluded that it is safe to assume (?) that the request seems to reach the TM1 server, and that TM1 returns something back (or simply closes the connection as suggested in the linkedin-post that I referenced in my first edit) . And hence, that it is safe to assume that this is something that have to be fixed on the BI server side(?).

The question is therefore more likely: Is it possible to change the BI server behavior so that it does not shut down when the TM1 server returns something invalid or closes the connection, and rather show some kind of error message instead?

Thanks for any input!

No correct solution

Licensed under: CC-BY-SA with attribution
Not affiliated with StackOverflow
scroll top