Timeout for content

Question

Hello
9iAS with Portal 9.0.2 in Sun Solaris
I am working with Oracle Reports Builder and Oracle Portal.
I've saved my reports with JSP extension.
So, I've made some portlets with these reports into PEOPLE_APP db provider.
Now, when I run my pages with these portlets, some times (most of times) the following error appears instead of the report:
Timeout for content=[any_number]
I've already researched in OTN and Metalink forums, I've changed my zone.properties and http.conf files, but it didn't resolve my problem.
So, what else I need to do in order to solve my problem? Any other ideas?
Thanks.
P.S: We've sent lots of questions with no satisfactory response from Oracle.
We've noticed that Oracle provided extremely poor support.
Maybe this isn't a good solution. Maybe there isn't a solution, at least in this way.
Take a look in some fragments of messages, found in OTN and Metalink forums:
"I've increased the timeout but that doesn't help.
It's also very strange that after rebooting the server in Portal 3.0.8/ias1.0.2.1 my pages appeared in 10-20 seconds and now in Portal 3.0.9/ias1.0.2.2 it takes two to four minutes. And this slow performance is only while retrieving the page for the first time. The second time I retrieve a page it's there in a few seconds.
So, I do not think it has anything to do with the provider's timeout...."
"But maybe it's the Portal-schema itself which causes this problem. I heard someone saying that he missed indexes after a Portal-upgrade --> maybe this is my problem too, although i didn't have any error-messages in the upgrade-logfiles...."
"Unfortunately we did not yet solve these timeout problems. Oracle is currently studying several logfiles and traces.
At first we thought that we had the timeout problems only after starting up the server. But now we noticed that the timeouts are repeated after 5 to 7 days. For some reason it looks as if Portal is refreshing the cache at that moment, but we can not find anywhere a refreshperiod of 120 hours or so..."
"Having wrestled with these Timeout errors for months with no satisfactory response from Oracle, I came across someone's solution yesterday and, for me, it worked immediately. Please keep in mind before doing this that it is in direct opposition to what the new "Troubleshooting Issues in Oracle9iAS Portal Architecture and Performance (3.0.9.8.2)" whitepaper states....."
"After many many performance problems and timeouts issues which Oracle provided extremely poor support we finally learned that all of our problems where related to the same things - Oracle has not tuned this product before shipping it. Numerous little bugs and coding errors were a small part of the fixes. The real problem is that all the indexes have subscriber id as the first column which is always set to one and is not used in the majority of querys which makes these indexes useless for performance... (http://metalink.oracle.com/metalink/plsql/ml2_gui.startup)"
"Sr. Oracle how long do I need to wait for real solutions? (http://metalink.oracle.com/metalink/plsql/ml2_gui.startup)"

Answer 1

You must have registered some provider for your portlets. Go to adminstration screens of portal. Goto edit provider from there and increase the portlet timeout.
Also change the time out value in your provider.xml for the corresponding JSPs.