We have a interesting problem that is causing MR (and IA) Reports to time out around 600 seconds (10 mins). However, when I execute the same FEX from the browser on the Client/Web Server (via RDC session), it doesn't time out.

Here is a example of the FEX from the Reporting Server that we are using to evaluate and test (thru the Portal) :

-*new file
-* pgm to determine how long browser times out
-SET &START = HHMMSS('A8');
-SET &CPUS= &FOCCPU;
-*-RUN
-* 2000 is 33 minutes, it crashes at about 10 min
-*SLEEP 2000
 
-* 600 is 10 min
SLEEP 600
 
APP PATH IBISAMP
TABLE FILE CAR
PRINT CAR
BY COUNTRY
HEADING
"THIS IS A TEST "
-*"START     &START    &CPUS "
-*"END  &END &CPUE" 
END
-RUN
 
-SET &END = HHMMSS('A8');
-SET &CPUE= &FOCCPU;
-TYPE START &START &CPUS
-TYPE END     &END     &CPUE
  

We have tested this FEX with using 'sleep 1500' and it doesn't time out on the server but 'sleep 600' does timeout on the desktop PC.

Any ideas to things to check?

This message has been edited. Last edited by: globalwm,

First off, why is it taking 10+ minutes to return your results?

There is, I think, a 10 minute timout setting, that is, if the browser does not receive results in 10 minutes, then it times out. This setting can be alter by your admin...

In normal use, in IA, when a user drags a column over without a WHERE clause on a Table with 25m rows, it takes > 10 as it does "Live Data" Preview.

To simulate this, we are using a simple FEX with a SLEEP cmd in it. This also rules out the RS/ORCL interaction as everything is on the RS.

The browser on the Client/Web server doesn't seem to have this time out as a successful test using 'SLEEP 3000' (50 minutes) just returned results.

Problem was on a F5 Protocol Client and Server Profiles: tcp_wan_optimized and tcp_lan_optimized were set at 300/600. Increased to 1800.