Re: Oracle 9.2.0 Gentoo Linux - ORA 3113 on DB startup

This is part of an strace prior to the ORA-3113 :

• SIGALRM (Alarm clock) --- rt_sigprocmask(SIG_BLOCK, [], NULL, 8) = 0 gettimeofday({1044994534, 354733}, NULL) = 0 rt_sigprocmask(SIG_UNBLOCK, [], NULL, 8) = 0 rt_sigprocmask(SIG_SETMASK, [RTMIN], NULL, 8) = 0 rt_sigprocmask(SIG_BLOCK, [ALRM], NULL, 8) = 0 setitimer(ITIMER_REAL, {it_interval={0, 0}, it_value={0, 0}}, NULL) = 0 rt_sigprocmask(SIG_UNBLOCK, [ALRM], NULL, 8) = 0 gettimeofday({1044994534, 355056}, {4294967236, 0}) = 0 shmdt(0x50000000) = 0 shmget(2268058976, 0, 0) = 1609596931 shmctl(1609596931, IPC_STAT, 0xbfffc814) = 0 shmat(1609596931, 0, 0) = 0x42659000
• SIGSEGV (Segmentation fault) --- rt_sigprocmask(SIG_UNBLOCK, [SEGV], NULL, 8) = 0

As you can see it segfaults after the shmat. Maybe somebody with a working installation could do an strace on the
oracle server process, so I can check what the differences between the 2 traces are...

"Jasper" <rig0r_at_phreaker.net> wrote in message news:3e495971$0$57460$1b62eedf_at_news.euronet.nl...
> Thanks. Still can't get it to work btw, I've tried all tips & tricks I
could
> find on this topic.
>
> I think I'll take another look at the stack trace.
>
> "anfo" <root_at_nycap.rr.com> wrote in message
> news:1jY1a.32784$Xr1.7562447_at_twister.nyroc.rr.com...
> > I am using the RedHat7.2 distribution default smp kernel. The system is
a
> > Dell 6300 quad 550-2mb xeon/4GB RAM/16GB Raid 5 array (s/w & h/w are on
> > "approved" list). None of these suggestions worked for me so far, I
> > increased shmmax from 1GB to 2 GB, didnt help initially. Once I took
down
> > the Running instance and the Orion appserver (an early 9iAS from
> ironflare)
> > the system still had a large amount of shared mem allocated to oracle,
> after
> > I manually removed the segments, I was able to create the new instance
> with
> > no problems. This did get me thinking about the Java implementation as I
> > noticed almost every oracle tool/utility is java based. I have the
> blackdown
> > jre1.1.8_v3, I am wondering if using the same 1.3.1_05 version that the
> app
> > server is using might help. I am also going to update the kernel to the
> > latest stable branch and build a custom, Ill try this next weekend and
> post
> > back if I find anything useful.
> >
> > "Howard J. Rogers" <howardjr2000_at_yahoo.com.au> wrote in message
> > news:pan.2003.02.10.21.54.25.384358_at_yahoo.com.au...
> > > On Mon, 10 Feb 2003 22:13:12 +0100, Jasper wrote:
> > >
> > > > Thanks for the answers guys.
> > > >
> > > > My answers :
> > > >
> > > > I've checked <all> kernel parameters. Rebooting would not be
necessary
> > > > because I
> > > > can write the new params in /proc/sys/kernel and they will be
> effective
> > > > immediately.
> > > >
> > > > Also tried the orarun9i script, no luck either.
> > > >
> > > > I've fiddled around with my init.ora, increasing/decreasing sga
size,
> > > > processes, etc, no luck.
> > > >
> > > > I'm using kernel 2.4.20 btw, what kernel version are you guys using
?
> > > >
> > >
> > > It works fine (sort of... see posts elsewhere for minor niggles) with
> > > 2.4.18 (Red Hat 8.0) and 2.4.19 (Mandrake). Certainly no 3113's.
> > >
> > > With Red Hat, I had to make a minor adjustment that I'd never come
> across
> > > before (but that's only because I was a Windows user until the
> weekend!):
> > > namely, edit /etc/sysctl.conf and add the lines
> > >
> > > # Shared memory limit
> > > kernel.shmmax = 1073741824
> > >
> > > I know you're not using Red Hat, but you might find
> > > http://thomasfly.com/Oranux/Oracle9i.htm helpful... I did on Mandrake,
> for
> > > example. And it's where I got the sysctl.conf tip from, which made a
> > > difference.
> > >
> > > Regards
> > > HJR
> > >
> > >
> > >
> >
> >
>
>
Received on Tue Feb 11 2003 - 14:24:28 CST