RE: Help - Oracle Reports 9ias Rel 2 is hanging - update

For anyone interested: I found some useful things:

-found parameter engineResponseTimeOut was introduced to timeout long running or hung reports that will cause the entire report server to hang you will see this below

-flakiness of separate report server processes versus in-process report server

-setup redundant report servers for same code and database=20

=20

Turned on Tracing

Editing the configuration file for reports over em the configuration file had ^M on every line which you see sometimes if you upload a file generated in the Windows environment and which of course would not work. Based on this behaviour, I recommend we copy the reports configuration that is working then edit it in vi on Unix rather then from the EM console. =20

=20

Also, I turned tracing on and explicitly directed trace files to /dblogs/9ias instead of the default. Even Oracle responds to different report issue threads by asking you to turn on this trace and sending it to them. Then added reports server as a new target to EM and we can monitor this trace from the EM console - hoping for some valuable information.

engineResponseTimeOut

The issue with this is to (1) allow for long-running reports if needed while (2) allow for hung reports to be "cleared" without having to cycle the reports server. I found that a parameter called engineResponseTimeOut was introduced to handle the situation when a long running or problematic report causes the report server engine to hang. Indicating that reports can hang the report server. I did not change the configuration yet to add this parameter as (1) I don't know what the outside limit should be. When the report server hangs that you see one report running and other reports "queing up". We should be able to see this from the EM console.

PROBLEM DESCRIPTION=20

--------------------=20

On occasion a report will hang the reports server engine.=20

You are not able to kill the report from the Queue manager =20

(using rwservlet/killjobid<id>?...)=20

=20

Any subsequent requests sent to the server just get queued behind the job you =20

cannot kill.=20

=20

The only way of clearing the hanging job is by bringing down the whole enterprise =20

edition (Middle Tier first and then the Infrastructure) and then reboot the server.=20

=20

=20

SOLUTION =20 ---------=20

 =20

A new attribute was added with Patch 1 to terminate hung or long running reports.=20

=20

"engineResponseTimeOut", has been added to the engine element in the Reports =20

Server configuration file.=20

=20

This attribute enables you to specify the maximum amount of time (in minutes) =20

for an engine to update the status of the job while running a report in your =20

environment. =20

If it takes longer than this amount of time to update the job status for some =20

reason (e.g., due to the engine hanging or a long blocking SQL query), =20

the Reports Server terminates the job.=20

=20

Sample engine tag to terminate jobs after 10 mintes will be : =20

=20

<engine id=3D"rwEng" class=3D"oracle.reports.engine.EngineImpl" initEngine=3D"1" =20

maxEngine=3D"5" minEngine=3D"0" engLife=3D"1" maxIdle=3D"30" callbackTimeOut=3D"60000" =20

engineResponseTimeOut=3D"10"> =20

<!--property name=3D"sourceDir" value=3D"/reports"/--> =20

<!--property name=3D"tempDir" value=3D"your reports temp directory"/--> =

</engine> =20

=20

NOTE : This is a sample tag, you may set the values according to your setup. =20

=20

Restart the Reports Server and OC4J_BI_FORMS instance.=20

=20

THE IN-PROCESS SOLUTION WAS SUGGESTED. This could be done by configuring the rep_<hostname>.conf and pointing the other configuration files to reference this instead of generating and using separate report processes, configurations. This could be easily reconfigured but would limit us to one report server per host.

=20

Another option configuring report services failover:

=20

Would have to be researched and tested. It can be used for failover and load-balancing. However, will it even work in this situation? As we saw when the reports server hangs the reports server process looks good, it's status is fine. Even with failover or load-balancing some reports will continue to work but will Oracle reports recognize a "hung" report server process and redirect to another report server or does the entire report server have to fail before redirection occurs? =20

-----Original Message-----
From: oracle-l-bounce_at_freelists.org
[mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Stankus, Paula G Sent: Friday, September 24, 2004 2:04 PM To: oracle-l_at_freelists.org
Subject: Help - Oracle Reports 9ias Rel 2 is hanging

Is there a way to get details on Oracle reports hanging? Has anyone else dealt with Oracle reports hanging? Is this a known issue? We are about to go to production with a system and the users have to run a batch process to cycle Oracle reports periodically. Is there anyway to trace the problem and determine if it is a problem with the code in the database that generates the error?

Thanks,
Paula

--
http://www.freelists.org/webpage/oracle-l


BEGIN-ANTISPAM-VOTING-LINKS
------------------------------------------------------
Teach CanIt if this mail (ID 11071242) is spam:
Spam:
https://dohsmsi01.doh.state.fl.us/canit/b.php?c=3Ds&i=3D11071242&m=3Ddd6f=
3fd6b
978
Not spam:
https://dohsmsi01.doh.state.fl.us/canit/b.php?c=3Dn&i=3D11071242&m=3Ddd6f=
3fd6b
978
Forget vote:
https://dohsmsi01.doh.state.fl.us/canit/b.php?c=3Df&i=3D11071242&m=3Ddd6f=
3fd6b
978
------------------------------------------------------
END-ANTISPAM-VOTING-LINKS

--
http://www.freelists.org/webpage/oracle-l