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

Re: [Xen-users] Me vs. RHEL 2.6.18-53.1.4.el5xen



The installation is up but the clients have crashed several times.

Here's the kernel dump on the console.

fg6x2.fnal.gov login: BUG: unable to handle kernel NULL pointer dereference at virtual address 00000670
 printing eip:
c0456a7c
2c438000 -> *pde = 00000003:2be75027
2c439000 -> *pme = 00000000:00000000
Oops: 0000 [#1]
SMP
last sysfs file: /class/net/virbr0/address
Modules linked in: ipt_MASQUERADE ipt_REJECT bridge ipv6 autofs4 i2c_dev i2c_core sunrpc xennet ip_conntrack_netbios_ns xt_state iptable_filter ipt_REDIRECT xt_tcpudp iptable_nat ip_nat ip_conntrack nfnetlink ip_tables x_tables dm_mirror dm_mod parport_pc lp parport pcspkr usb_storage megaraid_sas sd_mod scsi_mod ext3jbd ehci_hcd ohci_hcd uhci_hcd xenblk
CPU:    0
EIP:    0061:[<c0456a7c>]    Not tainted VLI
EFLAGS: 00010202   (2.6.18-53.1.4.el5xen #1)
EIP is at __handle_mm_fault+0x292/0x1070
eax: 00000670   ebx: 00000000   ecx: 00000000   edx: 00000670
esi: 000e5ebe   edi: 00000000   ebp: c1000000   esp: eca81e30
ds: 007b   es: 007b   ss: 0069
Process dd (pid: 2547, ti=eca81000 task=c0ee3000 task.ti=eca81000)
Stack: f57b4000 00000200 09cce004 ec423804 c356f200 00000000 00000000 00000000 00000000 00000000 00000000 ec439270 00000670 ec78a1e8 eca81ea0 c0406dad 00000000 00000000 00000000 c0e26d04 c356f248 00000670 ffffffff ffffffff
Call Trace:
 [<c0406dad>] do_IRQ+0xa5/0xae
 [<c0406dd2>] do_softirq+0x1c/0xae
 [<c05fdb48>] do_page_fault+0x0/0xbeb
 [<c0405515>] hypervisor_callback+0x3d/0x48
 [<c05fdb48>] do_page_fault+0x0/0xbeb
 [<c05fe230>] do_page_fault+0x6e8/0xbeb
 [<c045926d>] __vma_link+0x5b/0x68
 [<c0459a4b>] vma_link+0x4a/0xb8
 [<c0459ed9>] do_brk+0x1a8/0x1dc
 [<c043f4a4>] audit_syscall_exit+0x2b8/0x2d4
 [<c05fdb48>] do_page_fault+0x0/0xbeb
 [<c04054d3>] error_code+0x2b/0x30
 =======================
Code: 89 f0 c1 e0 05 8d 44 05 00 e8 01 bb ff ff 8b 54 24 08 c1 ea 09 81 e2 f8 0f 00 00 01 d0 89 54 24 54 89 44 24 30 0f 84 72 0d 00 00 <8b> 08 8b 68 04 f6 c1 81 89 8c 24 e0 00 00 00 0f 85 5d 0c 00 00
EIP: [<c0456a7c>] __handle_mm_fault+0x292/0x1070 SS:ESP 0069:eca81e30
 <0>Kernel panic - not syncing: Fatal exception

-----------------------------------------------------
Any idea what is wrong?  Is this something that can
be worked around or should I be thinking about of backing out?

Steve Timm




------------------------------------------------------------------
Steven C. Timm, Ph.D  (630) 840-8525
timm@xxxxxxxx  http://home.fnal.gov/~timm/
Fermilab Computing Division, Scientific Computing Facilities,
Grid Facilities Department, FermiGrid Services Group, Assistant Group Leader.

On Wed, 12 Dec 2007, Steven Timm wrote:

Now we see the network, thanks very much.  Up with
dom0 = 2.6.18-53.1.4.el5xen 64-bit and domu = 2.6.18-53.1.4.el5xen 32-bit.

Steve


------------------------------------------------------------------
Steven C. Timm, Ph.D  (630) 840-8525
timm@xxxxxxxx  http://home.fnal.gov/~timm/
Fermilab Computing Division, Scientific Computing Facilities,
Grid Facilities Department, FermiGrid Services Group, Assistant Group Leader.

On Wed, 12 Dec 2007, Steven Timm wrote:

On the first try of changing the xen.cfg to use xvda rather than
sda, I forgot to change the "root" clause to also point to xvda.
Now my domain is up under el5xen kernel but it has no network--
nevertheless this is progress.

Still, does anyone know--under red hat should I be using the .el5xen
or the .el5PAE kernel for the guest?

Thanks

Steve Timm



On Wed, 12 Dec 2007, Steven Timm wrote:

On Wed, 12 Dec 2007, Sadique Puthen wrote:


How did you create your initrd image?

Recreate it using --preload=xenblk --with-xennet. Change the disk line as below.

disk = [ 'phy:LG0/LV1,xvda,w','phy:LG0/LV2,xvdb,w','phy:LG0/LV3,xvdc,w' ]

Then try to start the guest.

--Sadique

I did not have these options in the initial initrd.  I remade the
initrd for 2.6.18-53.1.4.el5xen and got the same error as before,
it couldn't find the /dev/root.

Should I have done the 2.6.18-53.1.4.el5PAE kernel instead?

Does it matter that I am doing the mkinitrd while in chroot
to a 32-bit distro, but still actually running a 64-bit kernel?
Is the 3ware controller a problem here? (I presume that I was supposed to change the /etc/fstab of the
FS of the xen instance as well, I did so).

I tried to change the other .cfg file with the older
xensource kernel to use xvda, xvdb, etc and now that won't start
either.  I get the error "device xvda (51712,vbd) is already connected),
due, no doubt, to earlier things which tried to use the device
crashing and not closing it properly.  Any way to get around that?

Thanks

Steve Timm




Steven Timm wrote:
<flxi05> grep -v ^# xen1.cfg

kernel = "/linux32/boot/vmlinuz-2.6.18-53.1.4.el5xen"
ramdisk = "/linux32/boot/initrd-2.6.18-53.1.4.el5xen.img"
memory = 4000
name = "fcdfx41"
vif = [ 'mac=00:16:3e:0C:04:01, bridge=xenbr0', 'mac=00:16:3e:0C:04:09, bridge=xenbr1' ] disk = [ 'phy:LG0/LV1,sda1,w','phy:LG0/LV2,sda2,w','phy:LG0/LV3,sda3,w' ]
netmask= "255.255.255.0"
gateway= "131.225.240.200"
hostname= "fcdfx41.fnal.gov"
root = "/dev/sda1 ro"
extra = "4"

----------------
I am using the kernel inside dom0 to try to boot the domU.  That is
what I have done on all my installations up until now.  But I am willing
to change if someone finds a better way.

I also tried to install a different Xen instance from scratch using
virt-manager, but it did get up to installing about ten rpms before it
quit, i.e. it got through the whole anaconda input phase of the install.

Steve Timm





------------------------------------------------------------------
Steven C. Timm, Ph.D  (630) 840-8525
timm@xxxxxxxx  http://home.fnal.gov/~timm/
Fermilab Computing Division, Scientific Computing Facilities,
Grid Facilities Department, FermiGrid Services Group, Assistant Group Leader.

On Wed, 12 Dec 2007, Sadique Puthen wrote:


Can you please show your guest configuration file? Are you using a kernel and ramdisk from the dom0 to boot the guest or are they put inside guest filesystem and using pygrub to load them from guest?

--Sadique

Steven Timm wrote:

I am trying to migrate an existing working xen setup
which was based on xen 3.1.0 tarballs as they were available
at xen.org.  Underlying distro is RHEL5.
I am now trying to use the not-really-xen3.1.0-but-almost
xen that comes with RHEL5 update 1.

Host was 64-bit dom0. first domU was PAE 32-bit, both from xen 3.1.0 tarballs.

Now I have installed kernel-2.6.18-53.1.4.el5xen on the dom0.
That boots up fine.  I tried to start up the domU first with
the old 2.6.18-xen kernel (32-bit PAE) that was working before.  That
also started up OK, although it did crash once.

The question is--how to get the 32-bit PAE kernel from
RedHAt installed and booting in the domU.  I tried both
2.6.18-53.1.4.el5xen (32bit) and 2.6.18-53.1.4.el5PAE (32bit)

the former cannot find the root system file system of the domU. I get the following error:

mount: could not find filesystem '/dev/root'
Setting up other filesystems.
Setting up new root fs
setuproot: moving /dev failed: No such file or directory
no fstab.sys, mounting internal defaults
setuproot: error mounting /proc: No such file or directory
setuproot: error mounting /sys: No such file or directory
Switching to new root and running init.
unmounting old /dev
unmounting old /proc
unmounting old /sys
switchroot: mount failed: No such file or directory
Kernel panic - not syncing: Attempted to kill init
-----------------

If I compare between the 2.6.18-xen kernel from xensource which
works, and this one that doesn't, I see that I am missing
the modules ide-disk.ko, libusual.ko, usbcore.ko, usb-storage.ko
The usb modules should be irrelevant since there is no usb on this machine.
alas, the ide-disk.ko module which seems to be the key one isn't even
there at all in the redhat ersion of the kernel.

----------------------

If I try to load the PAE version of the kernel instead, I get the following error message:

VmError: (2, 'Invalid kernel', 'xc_dom_parse_elf_kernel: ELF image has no shstrtab\n')


Any clue what may be going on?
Does anyone else have a RHEL 5 update 1 system working with 64-bit host,
32 bit PAE guests?

Thanks

Steve Timm




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






--
------------------------------------------------------------------
Steven C. Timm, Ph.D  (630) 840-8525
timm@xxxxxxxx  http://home.fnal.gov/~timm/
Fermilab Computing Division, Scientific Computing Facilities,
Grid Facilities Department, FermiGrid Services Group, Assistant Group Leader.


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


_______________________________________________
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®.