[Info-Ingres] procedure create crashes out of server.
Martin Bowes
Hi All,
Does anyone recognie this?
In Ingres II 9.1.1 (a64.lnx/103)NPTL +p13001
CREATE PROCEDURE mb_flag_duplicate AS
BEGIN
UPDATE next_exceptionid
FROM
(
SELECT max(new_kuid) new_kuid
FROM kuids
) highest_kuid
SET exceptionid = char(numeric(highest_kuid.new_kuid+5));
END;
Executing . . .
E_RD0125 consistency check
- unexpected exception in RDF
(Wed Nov 4 13:07:16 2009)
E_SC0206 An internal error prevents further processing of this query.
Associated error messages which provide more detailed information about
the problem can be found in the error log, II_CONFIG:errlog.log
(Wed Nov 4 13:07:16 2009)
Exiting session because of communications failure.
In the errlog a stack trace is generated:
-----------BEGIN STACK TRACE------------
0: 4065f3e0 iidbms(rdf_readtuples+0xc9) [0x4b107d]( ... )
1: 406603c0 iidbms(rdf_call+0x2e8) [0x4a60d3]( ... )
2: 40660ae0 iidbms [0x54bc7c] <static function, no info available>
3: 406616f0 iidbms [0x54a044] <static function, no info available>
4: 40661f60 iidbms(psy_protect+0x674) [0x548364]( ... )
5: 40663ad0 iidbms(psy_view+0x1e50) [0x556cc0]( ... )
6: 40663b60 iidbms(psy_qrymod+0x375) [0x7d196d]( ... )
7: 40663f00 iidbms(psls39iftn+0x12b6) [0x7c6c96]( ... )
8: 40666980 iidbms(pslsparse+0xa0e) [0x792251]( ... )
9: 40666ba0 iidbms(psq_parseqry+0x503) [0x522203]( ... )
10: 40666dc0 iidbms(psq_call+0x2d5) [0x51c9a5]( ... )
11: 40668fa0 iidbms(scs_sequencer+0x265e) [0x479eee]( ... )
12: 40671120 iidbms(CSMT_setup+0x4e0) [0x966340]( ... )
Martin Bowes
Paul Mason
Could be bug 121807 - whose description mentions creating a DBP referencing a table that doesn't exist, but I bet could fail on this too.
I tested on l&g 9.1.2 using char(10)'s for the relevant columns and it created the DBP without complaining.
If you want to send me the copy.in for the real tables I'd be happy to check that too.
Cheers
Paul
Ingres Forums
Hello Martin,
E_RD0125 yes, but stack does not tell me anything.
Testcase seems to be easy. Could you provide table structure, so I can
check?
Bilgilhan
--
bilgihan
------------------------------------------------------------------------
bilgihan's Profile: http://community.ingres.com/forum/member.php?userid=8658
View this thread: http://community.ingres.com/forum/showthread.php?t=11313
Ian Kirkham
Hi Marty,
This is a scalar subquery – you’ll need to wait just a little bit longer for this.
Regards,
Ian
Karl Schendel
On Nov 4, 2009, at 1:57 PM, Ian Kirkham wrote:
>
> Hi Marty,
> This is a scalar subquery � you�ll need to wait just a little bit
> longer for this.
But the kind in a FROM clause ought to be OK in 9.1.1. Might "just"
be a bug.
Karl
> _______________________________________________
> Info-Ingres mailing list
> Info-...@kettleriverconsulting.com
> http://ext-cando.kettleriverconsulting.com/mailman/listinfo/info-
> ingres
Paul Mason
On Nov 4, 2009, at 1:57 PM, Ian Kirkham wrote:Hi Marty, This is a scalar subquery � you�ll need to wait just a little bit longer for this.But the kind in a FROM clause ought to be OK in 9.1.1. Might "just" be a bug. Karl
As I said, I managed to create (and run) the DBP without any problems on a 9.1.2 codeline with the latest fixes.
Paul Mason
Marty,
Below is the output from my 9.1.2 build area – as you can see it appears to be working.
So it looks like it’s been fixed. I would suspect bug 121807 as I said – because it’s in the right area and because it’s not fixed in the patch you refer to.
Have you tested it against the latest patch on ESD?
Cheers
Paul
INGRES TERMINAL MONITOR Copyright 2007 Ingres Corporation
Ingres 2006 Release 2 Linux Version II 9.1.2 (int.lnx/100)NPTL login
Thu Nov 5 12:25:39 2009
continue
*
CREATE PROCEDURE flag_duplicate AS
BEGIN
UPDATE next_exceptionid
FROM
(
SELECT max(new_kuid) new_kuid
FROM kuids
) highest_kuid
SET exceptionid = char(numeric(highest_kuid.new_kuid+5));
END;
* * * * * * * * * * * * \g
Executing . . .
continue
* select * from next_exceptionid\g
Executing . . .
┌────────────┐
│exceptionid │
├────────────┤
│990100000016│
└────────────┘
(1 row)
continue
* execute procedure flag_duplicate\g
Executing . . .
continue
* select * from next_exceptionid\g
Executing . . .
┌────────────┐
│exceptionid │
├────────────┤
│990100000015│
└────────────┘
(1 row)
continue
* select * from kuids\g
Executing . . .
┌────────────┬────────────┐
│old_kuid │new_kuid │
├────────────┼────────────┤
│12000v6kfj5j│990100000000│
│120016u79o6e│990100000005│
│12001a4g66du│990100000010│
└────────────┴────────────┘
(3 rows)
continue
* \q
Your SQL statement(s) have been committed.
Ingres 2006 Release 2 Version II 9.1.2 (int.lnx/100)NPTL logout
Thu Nov 5 12:27:10 2009
From: info-ingr...@kettleriverconsulting.com [mailto:info-ingr...@kettleriverconsulting.com] On Behalf Of Martin Bowes
Sent: 05 November 2009 08:58
To: Ingres and related product discussion forum
Subject: Re: [Info-Ingres] procedure create crashes out of server.
Hi Paul,
Sorry for the delay, I've been running around like a cut cat lately.
I've attached a tar file of all the goodies you'll need to visit Grandma's house.
Marty
From:
info-ingr...@kettleriverconsulting.com [mailto:info-ingr...@kettleriverconsulting.com]
On Behalf Of Paul Mason
Sent: 04 November 2009 14:16
To: Ingres and related product discussion forum; Ingres and related
product discussion forum
Subject: Re: [Info-Ingres] procedure create crashes out of server.
Could be bug
121807 - whose description mentions creating a DBP referencing a table that
doesn't exist, but I bet could fail on this too.
I tested on l&g 9.1.2 using char(10)'s for the relevant columns and it
created the DBP without complaining.
If you want to send me the copy.in for the real tables I'd be happy to check
that too.
Cheers
Paul
-----Original Message-----
From: info-ingr...@kettleriverconsulting.com on behalf of Martin Bowes
Martin Bowes
Hi Paul,
I've got an agreement to upgrade the site to 9.2.0+p13556 in the next few weeks. So I'll head off down that route.