Oracle 7.3.2, Rogue Wave DBTools.h++, Multithreading and seg violation problem on AIX.

From: Arthur J. Lewis <arthurlewis_at_prodigy.net>
Date: 1998/03/13
Message-ID: <6ec6q6$8kri$1@newssvr03-int.news.prodigy.com>#1/1

I have an application that is should run multithreaded and that uses DBTools.h++ (2.2) with Oracle (7.3.2) on AIX.

I do call the OCI function opinit(1);

It blows up with the dbx trace below. Anyone else seen this. If I serialize access to DBTools.h++ and consequently Oracle with the front end running multithreaded - I run just fine.

Any help, suggestions would be gratefully appreciated since I am currently in a multithreaded hell until this is solved.

PS - As soon as we have Oracle (7.3.3) up, I willl pray to the threading gods and will try that.

dbx trace of where it dies.

---

Segmentation fault in bzero.bzero [/usr/lib/libc_r.a] at 0xd02cec58 ($t14) 0xd02cec58 (bzero+0x54) 7ca305aa stswi r5,r3,0x0
(dbx) where

bzero.bzero() at 0xd02cec58

kpumgs(??) at 0xd1052698
kpusc(??, ??, ??, ??, ??, ??, ??, ??) at 0xd10509b0
kpuscn(??, ??, ??, ??, ??, ??) at 0xd107eac8
kpulsc(??, ??, ??, ??, ??) at 0xd105139c

upiosd(0x203b8b7c, 0x1, 0x20649d84, 0x14f, 0x2, 0x1, 0x0, 0x203b8b94) at 0xd107c688
ocisq7(??, ??, ??, ??, ??) at 0xd104f6cc parse__23RWDBOracleCallInterfaceFiRC9RWCStringRC13RWDBOracleCDAP10RWDBStatus
(0x1, 0x2063d140, 0x20587fb4, 0x204e0754) at 0x10528ae4
parse__22RWDBOracleStatementImpFRC9RWCStringRC14RWDBConnectionR10RWDBStatus( 0x20587f18, 0x2063d140, 0x2032beac, 0x204e0754) at 0x10540ddc parse__19RWDBSelectorBaseImpFRC14RWDBConnectionR10RWDBStatus(0x2063d0a8, 0x204e074c, 0x2032beac, 0x204e0754) at 0x10144874 execute__19RWDBSelectorBaseImpFRC14RWDBConnection(0x2063d0a8, 0x204e07dc, 0x2032beac) at 0x10144c14
readerImp__19RWDBSelectorBaseImpFRC14RWDBConnection(0x2063d0a8, 0x2032beac) at 0x10143a6c
reader__16RWDBSelectorBaseCFRC14RWDBConnection(0x204e094c, 0x204e0a3c, 0x2032beac) at 0x1013e350
... rest of the call stack for the application has been snipped ...

Thanks to anyone out there who has any info or helpful comments on this.

1. Lewis

Received on Fri Mar 13 1998 - 00:00:00 CST