Ansicht
Dokumentation

03813 - SQL 502: Cursor already open

03813 - SQL 502: Cursor already open

General Data in Customer Master   Fill RESBD Structure from EBP Component Structure  
This documentation is copyright by SAP AG.
SAP E-Book

SQL 502: Cursor already open

KERNEL Patch for 4.5B:
----------------------
Last night, I attempted to load the latest kernel patch for release 4.5B
(735). This was planned in response to the fact that some 502 errors
disappeared after kernel patch 719 was applied by other listservers.
Unfortunately, there were errors opening a lock handler file with the new
kernel level.

The errors I received in SM21 were as follows:
Error opening a lock handler file
> /usr/sap/DV2/DVEBMGS10/log/ENQBC enxxmfil0186
> The value specified for the argu enxxmfil0186

I ran an enqueue diagnostic check and it completed without errors. However,
I didn't feel comfortable continuing with this kernel level until I
understood the problem better. So, I reverted back to our previous kernel
level (573).

SQL 502 Errors:
---------------
Back to the original problem...I think I've determined that any BDC call
from an application server will result in the 502 error. Originally, I
thought this user had been running the BDC sessions on our application
server for some time, however, I since discovered that they inadvertently
deleted their PRD entry from SAPLogon, and our PC group reloaded the wrong
file. That new SAPLogon logs the user onto our application server. We
received our first 502 short dump within minutes of when the user began
processing on the appl. server.

So a general question for the group: Has anyone been able to successfully
run a BDC program on an AS/400 application server? Anyone on 4.5B
specifically?

Volker: OSS went down when I was submitting the OSS message (All the
information was included and I was just reviewing 🙁 ). I still haven't
recreated the message, but I will if I have to. My guess is that if you run
a BDC session on any appl. server, you too can receive the 502 error. This
would allow you to debug on your servers rather than our PRD system (we only
have appl. server for our production system). Let me know how you would
like me to proceed.

BTW, great idea for the photo album...I will submit a mug-shot next week.
🙂

Thanks,
Mike D. Martin
SAP Basis Administrator
SOLA Optical, USA
707-763-9911 x6106
mmartinZs...


-----Original Message-----
From: Gueldenpfennig, Volker [mailto:volker.gueldenpfennigZs...]
Sent: Tuesday, November 06, 2001 10:02 AM
To: 'Mike Martin, IS-NA, AOSOLA'
Cc: Sap400 (E-mail)
Subject: RE: SQL 502: Cursor already open


Hi Mike,

first of all, you are fortunately on the latest DBSL-level, as after there
was no change any more.

How to determine an SQL0502 problem?
It would be the easiest, if you could reproduce this problem in a dia_WP.
Then you could kill the first4 WPs hard on OS level in a more or less
"quite" time. The you go to SE38 and start the report RSTRC000. Here you can
mark an "X" at "hold WP". When you then have a look in SM50 from another WP
you see it "locked". This means that you now occupy a WP (similar to
debugging, but more strange). Now you do a STRSRVJOB on OS-level against
this WP and issue then STRDBG UPDPROD(*YES). You change the job to *PRTWRAP,
so that all wrapped joblog-data becomes dumped to spool. It is necessary to
kill the WP before, because you should have as few ODPs opena possible,
because otherwise you perhaps won't see the cursor-names that cause SQL0502
twice in the joblog.
When you now issue in this WP your report, you will encounter the problem
and can search for this cursor if it was already opened before. Then you
should have a look to these entries if a commit is in between to the error
and if at least the new cursoris a "with hold" cursor or not. You can check
this in the dev-trace. Unfortunately we can't check this at the moment for
the old cursor.

I think your problem only occured when 3-tier, then you have to use the
DB-shadow-job with STRDBG and STRSRVJOB. You can find out the name and
number in the dev-trace.

My DBSL-collegue thinks about changing the cursor-naming-conventions, so
that we can see if this is a "with hold" cursor or not. Then the analysis
would become a lot easier.

When you opened an OSS ticket, let me know, I will take this over then.

Regards

Volker


[Non-text portions of this message have been removed]


Durban Tours - Südafrika Safari

Vendor Master (General Section)   RFUMSV00 - Advance Return for Tax on Sales/Purchases  
This documentation is copyright by SAP AG.

Length: 5101 Date: 20240427 Time: 024100     sap01-206 ( 3 ms )