Re: OAS report server sometimes does not create reports

From: Yong Huang <yong321_at_yahoo.com>
Date: Fri, 7 Sep 2012 14:58:49 -0700 (PDT)
Message-ID: <1347055129.27965.YahooMailClassic_at_web184803.mail.gq1.yahoo.com>



Dimitre,

Sorry for getting my report server version wrong. It's 10.1.2.3.0.

I've been running with report server and engine trace on: <trace traceOpts="trace_all"/>
for a few weeks. We only have less than 20 reports created on each report server every day. So the overhead of that is negligible. Unfortunately, even though the logs have some useful information e.g. java Exception stack and I suppose trace_all includes TRACE_PLS (thanks for bringing to my attention Note 237301.1), grep -i PKG_ACC_DATA or ORA-004020 does not appear in any log. I can see the report java process writes the deadlock info to a pipe, the read side of which is its parent, rwserver. It's not written to any file. Anyway, I don't know what to do with this deadlock that exists inside the reports engine only. Killing the process is a workaround.

Yong Huang

-----Original message-----

On Fri, 9/7/12, Radoulov, Dimitre <cichomitiko_at_gmail.com> wrote:

On 07/09/2012 19:39, Yong Huang wrote:
> Resend:
>
> Indeed, the PL/SQL engine can be in Oracle forms or reports. I wish
> it offered something equivalent to dba_* and v$* views, an easy way
> to monitor its working mechanism.
>
> For now, I'll write a cron job (in addition to another one that
> creates a dummy report) to run strace on all report server processes.
> If the output has strings like ORA-4020 or deadlock, it emails me.
>

Hi Yong,
you mentioned Oracle Application Server 10.1.0.3 but I suppose you mean 10.1.3. The following MOS Note could be helpful:

Logging and Tracing in Reports 9.0.x / 10.1.2.x [ID 237301.1]

I suppose that some of the settings provided in that document are valid for 10.1.3 as well. (there is another MOS Note but it's for 11g: Note:856664.1 Logging and Tracing in Reports 11g. A Quick Reference Guide).

Best regards

Dimitre

--
http://www.freelists.org/webpage/oracle-l
Received on Fri Sep 07 2012 - 16:58:49 CDT

Original text of this message