[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Xen-users] Re: Kernel Panic from dom0 with smart array 5i


  • To: xen-users@xxxxxxxxxxxxxxxxxxx, "Duquesne Olivier" <olivier.duquesne@xxxxxxxxxxxxxx>
  • From: DaffyDuke <daffyduke@xxxxxxxxx>
  • Date: Thu, 31 Jan 2008 17:53:25 +0100
  • Delivery-date: Thu, 31 Jan 2008 08:54:18 -0800
  • Domainkey-signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=message-id:date:from:to:subject:in-reply-to:mime-version:content-type:content-transfer-encoding:content-disposition:references; b=Gz5X93b5vt9IWNApUs/BvdUoQogti1hGQOQ7/ph95CVgizU7vjJGI4N/nTt5i+b0R9jraimGJpxM3tIY2A+o0ziCqgcNp5dLk3Ayqtj7zgONWHnV3c70PLZwR4saEloLv+fRBJRULQ0CtmKUvUf3BYZaLse/SUnJR+YmrupnxiU=
  • List-id: Xen user discussion <xen-users.lists.xensource.com>

Hi,

So I solved some of my problems :

On Jan 9, 2008 8:17 PM, DaffyDuke <daffyduke@xxxxxxxxx> wrote:
> Hi dear Xen users,
>
> We experienced kernel panic while acessing / filesystem with an HP DL-380 G3.
> cciss support is well activated , it is running with G2, G4 and G5 but
> not with G3.
>
> We tried to build from source, binary from Xen comunity and binary for
> RedHat, there's always same symptome.
> We tried Xen 3.01, 3.02, 3.05, Xen 3.1.
>
> HP CCISS is discovered but not the partition table, so, we can't mount
> /dev/cciss/c0d0p1 for / .
>
> Hardware is
> product: ProLiant DL380 G3
> description: RAID bus controller
> product: Smart Array 5i/532
> vendor: Compaq Computer Corporation
> Board ID: 0x40800e11


Indeed, it was fine since Xen 3.0.3. Solution was to remove acpi=off
from the grub line. We used this options to enable more than 4 Go of
RAM on other hardware.
Now, with Xen 3.2.0 (really a last checkout from mercurial, Xen 3.2.1
rc1 , just a few hours before the official announce on xen-devel
list), Xen starts on HP DL-380 G3 dual Xeon 2.8 GHz and 6 Gb RAM. dom0
and domU OK.
I had to change this on the domU configuration :
cpu is now vcpu
andd xencons=tty in the append line , console=tty on kernel and module
line from the grub.conf are not sufficient now.

Then,I made some brindging tests :
- NATIVE :  bridging eth0 <-> xenbr0 OK with network-bridge
- MULTIPLE NATIVE bridges in bridge.conf, KO with network-bridge. OK
with our bridge-wrapper, if we do not specified vif number from the
command line argument passed to network-bridge.
- VLAN : eth0.vlan bridged to xenbr.vlanid OK (need ONBOOT=yes for the
ifcfg-bond0.vlanid first or Xen doesn't initiate vconfig)
- BONDING : bond0 bonding on eth0+eth1 , bridged on xenbr0 : OK
- BONDING+VLAN : bond0.vlanid (over eth0+eth1) bridged on xenbr.vlanid
: KO => Kernel Panic.

My interface bond0.vlanid is up and running, but when Xend start, I
systematically failed with a kernel panic.

Here is some of my call trace :

cannot create /proc/net/bonding/<NULL>
BUG : unable to handle NULL pointer dereference
OOOPS : 0000 #1
EIP : 0061
EIP is at sysfs_remove_group
[...]
Call trace :
show_stack_log_lv
show_vg
die
do_page_fault
error_code 0x2b/0x30
bond_destroy_sysfs_entry [bonding]
bond_event_changename [bonding]
bond_master_netdev_event [bonding]
notifier_call_chain
dev_change_name
dev_ifctl
dev_ioctl
sock_ioctl
vfs_ioctl
syscall_call 0x7/0xb
EIP sysfs_remove_group 0x57/0xb0 SS ESP 0069 C0997E08


Hum .....    :-(
Sould anyone help me ?


Thanks a lot,

Daffy
http://www.linkedin.com/in/oduquesne

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.