Re: Most programs crash under openmosix.

Hi,

I actually use the oM user tools from the gentoo portage and as I have still
not yet set up a portage overlay for my own kernel build, the gentoo
openMosix kernel is intalled once (but not used) and this enables me to:
 emerge --buildpkg openmosix-user
on the 1st machine and then on all other machines I:
emerge --usepkg --nodeps openmosix-user
using the same package dir within make.conf over nfs.

Also note what Evan Hisey is asking as memory errors and even faulty network
connections can show up and be quite hard to trace expecially on old desktop
machines.

To test migration, i use Fremen's openMosix benchmark tool:
http://www.xyz-network.com/ombenchmark/
it uses perl and is very simple to use:

./mosixtest-<version> 10	# for 10 processes

Then monitor on the originating node with mtop and on any other node using
mosmon

Hope this helps and good luck

Martin



On 16/03/06, Gert Burger <[email protected]> wrote:
>
> THis actually makes sense so i will need to try it later. Which
> openmosix-user tools are you using?
>
> Thanks for replying.
>
> On Wed, 2006-03-15 at 20:53 +0000, Deadpan110 wrote:
> > Hi,
> >
> > I also use Gentoo and have tried a few of the openMosix kernels within
> > Gentoo portage.
> >
> > I am unsure of their current state but have always found them to be
> > unstable due to the amount of patches included with them (migshm and
> > many others).
> >
> > The official openMosix kernel is 2.4.26 and thus, I always build my
> > own and ignore the Gentoo ones all together.
> >
> > Martin
> >
> >
> >
> > On 15/03/06, Gert Burger <[email protected]> wrote:
> >         Hi
> >
> >         We run a smallish(18 pc) cluster for AI smulations. At least
> >         we are
> >         trying to.
> >
> >         The processes migrate properly but openmosixtest packages and
> >         almost all
> >         of our simulations crash after only a few seconds with a
> >         kernel oops.
> >         Below is the oops.
> >
> >         We are running openmosix 2.4.30 and openmosix-user 0.3.6-r3
> >         from the
> >         gentoo portage.
> >
> >         Any ideas anyone?
> >
> >         The openmosixtests freeze the machine within 5-10 seconds.
> >
> >         Oops from our program:
> >         <1>Unable to handle kernel NULL pointer dereference at virtual
> >         address
> >         0000004c
> >         c0172b9d
> >         *pde = 00000000
> >         Oops: 0002
> >         CPU:    0
> >         EIP:    0010:[<c0172b9d>]    Not tainted
> >         Using defaults from ksymoops -t elf32-i386 -a i386
> >         EFLAGS: 00010286
> >         eax: 00000000   ebx: ce7c4720   ecx: 0805672c   edx: 00000000
> >         esi: 08057000   edi: 08057000   ebp: 00000000   esp: cdfe1b2c
> >         ds: 0018   es: 0018   ss: 0018
> >         Process ls (pid: 1659, stackpage=cdfe1000)
> >         Stack: cdfe1b30 ce7c4660 ce7c4720 00000000 00000009 c017510d
> >         08056380
> >         0805672c
> >         00000380 00000003 00001812 0000000e c041f8f0 fffffffb 00000000
> >         00000246
> >         c041f8f0 00000246 00000002 000038c0 cf3651bc cf3650e0 000053aa
> >         00000380
> >         Call Trace:    [<c017510d>] [<c033336a>] [<c01748c0>]
> >         [<c0158624>]
> >         [<c015881f>]
> >         [<c0105da2>] [<c0109f3b>] [<c01f1aa7>] [<c0105d60>]
> >         [<c01f0613>]
> >         [<c01f6f5d>]
> >         [<c0109e16>]
> >         Code: 89 78 4c 89 78 48 8b 5c 24 08 31 c0 8b 74 24 0c 8b 7c 24
> >         10
> >
> >
> >         >>EIP; c0172b9d <set_brk+3d/c0>   <=====
> >
> >         Trace; c017510d <load_elf_binary+84d/10b0>
> >         Trace; c033336a <inet_recvmsg+4a/60>
> >         Trace; c01748c0 <load_elf_binary+0/10b0>
> >         Trace; c0158624 <search_binary_handler+154/200>
> >         Trace; c015881f <do_execve+14f/330>
> >         Trace; c0105da2 <sys_execve+42/80>
> >         Trace; c0109f3b <call_with_regs+4b/94>
> >         Trace; c01f1aa7 <deputy_syscall+a7/2b0>
> >         Trace; c0105d60 <sys_execve+0/80>
> >         Trace; c01f0613 <deputy_main_loop+433/590>
> >         Trace; c01f6f5d <mosix_pre_usermode_actions+6d/c0>
> >         Trace; c0109e16 <straight_to_mosix+5/d>
> >
> >         Code;  c0172b9d <set_brk+3d/c0>
> >         00000000 <_EIP>:
> >         Code;  c0172b9d <set_brk+3d/c0>   <=====
> >            0:   89 78 4c                  mov    %edi,0x4c(%eax)
> >         <=====
> >         Code;  c0172ba0 <set_brk+40/c0>
> >            3:   89 78 48                  mov    %edi,0x48(%eax)
> >         Code;  c0172ba3 <set_brk+43/c0>
> >            6:   8b 5c 24 08               mov    0x8(%esp),%ebx
> >         Code;  c0172ba7 <set_brk+47/c0>
> >            a:   31 c0                     xor    %eax,%eax
> >         Code;  c0172ba9 <set_brk+49/c0>
> >            c:   8b 74 24 0c               mov    0xc(%esp),%esi
> >         Code;  c0172bad <set_brk+4d/c0>
> >           10:   8b 7c 24 10               mov    0x10(%esp),%edi
> >
> >
> >
> >         -------------------------------------------------------
> >         This SF.Net email is sponsored by xPML, a groundbreaking
> >         scripting language
> >         that extends applications into web and mobile media. Attend
> >         the live webcast
> >         and join the prime developer group breaking into this new
> >         coding territory!
> >
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=110944&bid=241720&dat=121642
> >         _______________________________________________
> >         openMosix-general mailing list
> >         openMosix-general-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
> >         https://lists.sourceforge.net/lists/listinfo/openmosix-general
> >
>
>