Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Oracle Application Server Internal Server Error

Re: Oracle Application Server Internal Server Error

From: Frank van Bortel <frank.van.bortel_at_gmail.com>
Date: Tue, 30 Oct 2007 19:26:33 +0100
Message-ID: <fg7t0l$i1h$1@news5.zwoll1.ov.home.nl>


gnscon_at_gmail.com wrote:
> Hello,
>
> We are using OAS 9.0.4.1 run on AIX 5.3 with an Oracle 10.1.0.2 DB
> back end running on a seperate partition of the machine. The main
> application is developed in house on forms6 and migrated to forms9 a
> couple of years ago when we moved to AIX. Since the move we get an
> "Internal Server Error", in the web browser, occuring randomly when
> users try to login to the application.
>
> Through some investigation it appears as though the OC4J_BI_Forms
> service becomes unresponsive for a short period of time causing the
> web server request to time out. Users who are logged into the system
> can still work but no new users can get in. Has anybody else
> experienced this problem? Anybody got any ideas where to start
> looking?
>
> We have a TAR logged with Oracle but as usual they've said that they
> won't help us until we go to 9.0.4.3. Which we haven't been able to do
> as the patching process keeps failing.
>
> Any help would be appreciated.
>

Without the exact error stack, I can hardly tell whether I have seen *this* error. I have seen a load of internal Server errors, accompanied by FRM-9xxxx codes.

There are (several!) metalink documents on how to tackle these errors, start tracing, etc.
In the environment I witnessed these problems, it was due to process limitations on the OS (Windows); setting the "Allow service to Interact with the Desktop" option helped for some time - in the end the customer ended up with more iron - it now runs 6 machines, hardware load balancer.

I doubt the time out - although there is one particular flavor of the error stack that indeed records some http header time out. Not at work, so I can't check.

Anyway - the fastest road for you is to check metalink.

-- 
Regards,
Frank van Bortel

Top-posting is one way to shut me up...
Received on Tue Oct 30 2007 - 13:26:33 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US