Re: 10.2.0.2 Core dumps
Thanks for everyone's input..
catpatch has been run, and didn't report any errors.
ulimit settings are also there according to the docs.
Jesse, Rich <Rich.Jesse@qg.com> wrote:
curious: Does the output of "ulimit -a" show acceptable limits as described
in the installation guide? I've recently been shown the light that I've
been mistakenly skipping over that requirement, which has at least changed
between 9iR2 and 10gR2 for Linux (e.g. "ulimit -p").
Sent: Tuesday, May 16, 2006 1:57
Subject: 10.2.0.2 Core
Anyone else experience unusually
often occurring coredumps and segfaults on 10.2.0.2 (linux) ?
Even the upgrade from
10.2.0.1 to 10.2.0.2
core dumped on me on first run. After flashing the database back and retrying
it ran fine. However, now the main application running on this (test) system
caused oracle to dump core more frequently then not.
Getting loads of
different variants of ORA-7445 :
core dump [qesmmIUnRegisterWorkArea()+628]
[SIGILL] [Illegal operand][0x8EA33E0]  
[kglhpd_internal()+1298] [SIGSEGV] [Address not mapped to object][0x14C63C8]
core dump [kcbget()+1662] [SIGSEGV] [Address not mapped to object]
[0xE]  
core dump [eomg_migrate_contents_as_needed()+762] [SIGSEGV]
[Address not mapped to object] [0xFFF8]  
core dump [ksarcv()+328]
[SIGSEGV] [Address not mapped to object] [0x0]  
[kcbgcur()+2685] [SIGSEGV] [Address not mapped to object] [0xE] 
Anyone experience similar problems with this release