Re: [uml-user] Oracle 10 in UML, anyone?

[Konrad M. <ko...@su...>]

Hi Alex
Yup - had that too, although with Ora 9i.  As near as I can tell it has
to do with the invocation of the JVM by the installer.  I could even
cause a segfault with a simple "java -version"!

My host is running a fairly vanilla FC2 install, with network bridging.

As for the guests segfaulting, one problem was that I didn't have a
swapfile properly configured for the guest system.  

Also, I was trying to get Oracle installed on an FC1 filesystem with a
2.4 kernel, which just never worked, so I moved to a 2.6.6 kernel, which
was better, but didn't quite solve the problem.  What I did then was a
"mv /lib/tls /libdisabled-tls", which, astoundingly, sorted the problem!
Not sure whether other distros have tls, but this immediately solved the
problem for my FC1 filesystem.

I'm also told that downgrading the glibc to the i386 version from the
default i686 version would have had pretty much the same effect.

Interestingly enough, after the installation, I just cant run the
database creation agent: it keeps bombing out with some arbitrary memory
error now, which Google says has to do with the kernel settings, which
is rubbish, because I've set everything properly and it works on other
machines.  So sigh . . . just getting Oracle installed may not be the
end of your problems I'm afraid!

Later
Konrad


On Fri, 2004-07-23 at 15:21, Alexander Bochmann wrote:
> Hi,
> 
> I'm curious if anyone has ever tried running an 
> Oracle DB inside an UML...
> 
> Currently, I can't even complete the Oracle installer 
> (which I run in X (xvncserver) and which relies heavily 
> on java) without crashing the UML linux it's running in.
> 
> I'm using 2.4.26+skas3 on the host and the Debian/sarge 
> user-mode-linux package (based on 2.4.26-1um) for the 
> clients. Host and UML run Debian/sarge (yes, I know 
> that's not a supported OS for Oracle, but it works 
> without UML...).
> 
> As the debian package is stripped, and I didn't find an 
> accompanying System.map, I can't produce any meaningful 
> backtraces or such right now, but I'm currently building 
> an UML kernel with debug information, so that I'll perhaps
> have some more useful data about the crashes.
> 
> With three attempts I had the UML linux hanging once with:
> 
> >Kernel panic: Segfault with no mm
> >Kernel panic: kernel BUG at sched.c:557!
> >Kernel panic: kernel BUG at sched.c:557!
> [etc., until I killed the linux process]
> 
> Crash with:
> 
> >Kernel panic: Segfault with no mm
> >Kernel panic: Segfault with no mm
> >Kernel panic: Segfault with no mm
> >In idle task - not syncing
> 
> and another crash with:
> 
> >Kernel panic: Kernel mode fault at addr 0x3d5e3d8e, ip 0x400f62d8
> 
> Not very encouraging :)
> 
> Alex.
> 
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by BEA Weblogic Workshop
> FREE Java Enterprise J2EE developer tools!
> Get your free copy of BEA WebLogic Workshop 8.1 today.
> http://ads.osdn.com/?ad_id=4721&alloc_id=10040&op=click
> _______________________________________________
> User-mode-linux-user mailing list
> Use...@li...
> https://lists.sourceforge.net/lists/listinfo/user-mode-linux-user