DomU crashing on x86

DomU crashing on x86

[David F Barrera]

We are seeing a problem in xen-unstable where DomU crashes immediately
after booting on some x86 boxes. In my case, it is happening on a couple
of machines that have SCSI drives and where DomU is a SLES 9
installation. Here's some data on my setup:

		Dom0		DomU		Dom U Crash? 	SCSI

Machine A	SLES 9		SLES 9		Yes		Yes
Machine B	SLES 9 		RHEL 4		No		Yes
Machine C	RHEL 4		SLES 9		Yes		Yes
Machine D	FC3	  SLES 9 and Debian	No		No
						
The only thing I see in common in the machines that fail is the SCSI
drives.  Machines B and C are identical; the others are different
models.

This is the Oops info that I am seeing (no other info is displayed):

Oops:   0002 [#1]
PREEMPT SMP modules linked in:
CPU:    0
EIP:    0061:[<00244ec0>]       Not tainted VLI


David Wolinsky (david_wolinsky@dell.com) has also reported seeing the
problem:

------- Additional Comments From david_wolinsky@dell.com  2005-07-13
21:26 -------
I put together a testing suite to figure out the problem… and it seems
related 
to two parts of my machine.

A)  Fedora Core 3
B)  SATA drive (as you are probably aware these use the same scsi
driver)

To test A out, I installed Debian (known for its stability, etc)… And
ran with 
SATA in SATA mode and in IDE emulation mode (cominbation mode as listed
in the 
bios)… both tests passed (not concrete, I only ran it as long as I had
known 
the test of Fedora Core to last… I'll run it over night for 100%).  

To test B out, I set my SATA mode into IDE emulation mode and ran it in
FC3, 
this time it crashed the DomU and not Dom0.

My conclusion is... an issue between the SCSI driver, Xen hypervisor,
and the 
memory driver
-------------------end of David Wolinsky's comments ---------------

Has anyone else observed this problem?  I believe this is a serious
problem that requires some attention.

-- 
Regards,

David F Barrera
Linux Technology Center
Systems and Technology Group, IBM

"The wisest men follow their own direction. "
                                                        Euripides

Re: DomU crashing on x86

[Nivedita Singhvi]

David F Barrera wrote:
> We are seeing a problem in xen-unstable where DomU crashes immediately
> after booting on some x86 boxes. In my case, it is happening on a couple
> of machines that have SCSI drives and where DomU is a SLES 9
> installation. Here's some data on my setup:
> 
> 		Dom0		DomU		Dom U Crash? 	SCSI
> 
> Machine A	SLES 9		SLES 9		Yes		Yes
> Machine B	SLES 9 		RHEL 4		No		Yes
> Machine C	RHEL 4		SLES 9		Yes		Yes
> Machine D	FC3	  SLES 9 and Debian	No		No
> 						
> The only thing I see in common in the machines that fail is the SCSI
> drives.  Machines B and C are identical; the others are different
> models.

Dave, what are your boot options?

thanks,
Nivedita

> This is the Oops info that I am seeing (no other info is displayed):
> 
> Oops:   0002 [#1]
> PREEMPT SMP modules linked in:
> CPU:    0
> EIP:    0061:[<00244ec0>]       Not tainted VLI
> 
> 
> David Wolinsky (david_wolinsky@dell.com) has also reported seeing the
> problem:
> 
> ------- Additional Comments From david_wolinsky@dell.com  2005-07-13
> 21:26 -------
> I put together a testing suite to figure out the problem… and it seems
> related 
> to two parts of my machine.
> 
> A)  Fedora Core 3
> B)  SATA drive (as you are probably aware these use the same scsi
> driver)
> 
> To test A out, I installed Debian (known for its stability, etc)… And
> ran with 
> SATA in SATA mode and in IDE emulation mode (cominbation mode as listed
> in the 
> bios)… both tests passed (not concrete, I only ran it as long as I had
> known 
> the test of Fedora Core to last… I'll run it over night for 100%).  
> 
> To test B out, I set my SATA mode into IDE emulation mode and ran it in
> FC3, 
> this time it crashed the DomU and not Dom0.
> 
> My conclusion is... an issue between the SCSI driver, Xen hypervisor,
> and the 
> memory driver
> -------------------end of David Wolinsky's comments ---------------
> 
> Has anyone else observed this problem?  I believe this is a serious
> problem that requires some attention.
> 

Re: DomU crashing on x86

[David F Barrera]

On Wed, 2005-07-13 at 14:55 -0700, Nivedita Singhvi wrote:
> David F Barrera wrote:
> > We are seeing a problem in xen-unstable where DomU crashes immediately
> > after booting on some x86 boxes. In my case, it is happening on a couple
> > of machines that have SCSI drives and where DomU is a SLES 9
> > installation. Here's some data on my setup:
> > 
> > 		Dom0		DomU		Dom U Crash? 	SCSI
> > 
> > Machine A	SLES 9		SLES 9		Yes		Yes
> > Machine B	SLES 9 		RHEL 4		No		Yes
> > Machine C	RHEL 4		SLES 9		Yes		Yes
> > Machine D	FC3	  SLES 9 and Debian	No		No
> > 						
> > The only thing I see in common in the machines that fail is the SCSI
> > drives.  Machines B and C are identical; the others are different
> > models.
> 
> Dave, what are your boot options?

The boot options for DomU are:

# Kernel image file.
kernel = "/boot/vmlinuz-2.6-xenU"
# Optional ramdisk.
# ramdisk = "/boot/initrd-2.6.12-xenU"
# The domain build function. Default is 'linux'.
builder='linux'
# Initial memory allocation (in megabytes) for the new domain.
memory = 256
# A name for your domain. All domains must have different names.
name = "vm1"
disk = [ 'phy:sda9,0809,w','phy:sda8,0808,w' ]
# Set if you want dhcp to allocate the IP address.
# vif = [ 'mac= AA:00:00:47:CB:34, bridge=xen-br0' ]
# Set if you want dhcp to allocate the IP address.
#dhcp="dhcp"
# Set netmask.
#netmask = '255.255.255.0'
# Set default gateway.
# gateway = '9.53.94.0'
# Set the hostname.
# hostname= "vm%d" % vmid


# Set root device.
root = "/dev/sda9 ro"
# Sets runlevel 4.
# extra = "4"
#============================================================================

> thanks,
> Nivedita
> 
> > This is the Oops info that I am seeing (no other info is displayed):
> > 
> > Oops:   0002 [#1]
> > PREEMPT SMP modules linked in:
> > CPU:    0
> > EIP:    0061:[<00244ec0>]       Not tainted VLI
> > 
> > 
> > David Wolinsky (david_wolinsky@dell.com) has also reported seeing the
> > problem:
> > 
> > ------- Additional Comments From david_wolinsky@dell.com  2005-07-13
> > 21:26 -------
> > I put together a testing suite to figure out the problem… and it seems
> > related 
> > to two parts of my machine.
> > 
> > A)  Fedora Core 3
> > B)  SATA drive (as you are probably aware these use the same scsi
> > driver)
> > 
> > To test A out, I installed Debian (known for its stability, etc)… And
> > ran with 
> > SATA in SATA mode and in IDE emulation mode (cominbation mode as listed
> > in the 
> > bios)… both tests passed (not concrete, I only ran it as long as I had
> > known 
> > the test of Fedora Core to last… I'll run it over night for 100%).  
> > 
> > To test B out, I set my SATA mode into IDE emulation mode and ran it in
> > FC3, 
> > this time it crashed the DomU and not Dom0.
> > 
> > My conclusion is... an issue between the SCSI driver, Xen hypervisor,
> > and the 
> > memory driver
> > -------------------end of David Wolinsky's comments ---------------
> > 
> > Has anyone else observed this problem?  I believe this is a serious
> > problem that requires some attention.
> > 
> 
> 
-- 
Regards,

David F Barrera
Linux Technology Center
Systems and Technology Group, IBM

"The wisest men follow their own direction. "
                                                        Euripides