Urgent, DB2 ODBC Error SQLSTATE 57014

[David Grospelier]

Hi,

With DB2 5.2 on Unix and Windows NT 4 client with IBM ODBC Driver for
DB2 V5.02.00.00, we have this error when running a long query with
inner join and group by statement :
IBM CLI DRIVER DB2/6000 : Processing was cancelled due to an
interupt.
SQLSTATE=57014 (#-952)

The client run an Access 97 Application wich use the data to build
some reports. When we copy the data from the DB2 to Access local
tables, it runs correctly.
When we execute the query directly on the DB2, it runs correctly too.

We have made some traces and have, on the odbc trace, this error :

PREV 3f:5c EXIT SQLFetch with return code -1 (SQL_ERROR)
HSTMT 0x093f1830

DIAG [S1008] [IBM][CLI Driver][DB2/6000] SQL0952N Processing was
cancelled due to an interrupt. SQLSTATE=57014
(-952)

We have modified the APPLHEAPSZ parameter on the server, according to
newsgroup's messages but without success.

Many thanks for your help.

David Grospelier
dgros...@sword.fr

[PM (pm3iinc-nospam)]

i'd look for either cli/odbc settings that relate to time/timeout, ... or
maybe access settings.
(or maybe the governor?)

... and be sure the user doesn't intentionnaly disconnect/close access, ...

PM

"David Grospelier" <davidgr...@hotmail.com> a écrit dans le message de
news: 99a27579.02082...@posting.google.com...

[news.wanadoo.fr]

Hi,

I've tried to change Access settings relating to timeout and some others
things.
I'm sure the user doesn't intentionnaly close access because I'm the user !
:)

Thanks for help

David Grospelier
dgros...@sword.fr

"PM (pm3iinc-nospam)" <Pm3iinc...@sympatico.ca> a écrit dans le message
de news: mw5b9.5187$uF.8...@news20.bellglobal.com...

[sha...@us.ibm.com]

Hi,
I bet your client is sending a timeout interupt. The timeout can come from at least 5 places:
- user ctrl-c .... but you aren't doing that
- application (access)
- odbc driver
- db2 driver
- tcpip stack settings (keepalive too low?)

If you search the CLI Guide index for "time" it gives a number of ways to set timeouts from the odbc/cli driver issued interrupts.
Often, you can resolve this issue by turning the timeout value way up, to say 10 minutes.

If you want to get 'under the hood', take a cli trace with "tracecomm=1". If there is a timeout from the client side, you will see the corresponding network flow in the trace. Checkout the cli guide, appendix k, for tracing instructions.

Good luck,