On Sat, May 15, 2010 at 01:29:13PM +0800, lei yang wrote:
> Hi Pasi,
>
> I find some of my concept is not clear.
>
> After I boot domU, my domU root=/dev/sda1, I just have one hard disk and
> just one partition-->sda1
>
If using a pvops domU kernel then you shouldn't use sd* for the guest disk.
pvops kernels only support xvd* as a guest disk.
> and run
> [1]root@intel_5500_server:/root> service xend start &
> [1] 2114
> [2]root@intel_5500_server:/root> mount: xenfs already mounted or /proc/xen
> busy
> mount: according to mtab, none is already mounted on /proc/xen
> Nothing to flush.
Have you read:
http://wiki.xensource.com/xenwiki/XenParavirtOps
Especially the troubleshooting sections?
> [ 86.292484] xenbus_probe wake_waiting
> [ 86.296663] xenbus_probe wake_waiting
> Traceback (most [ 86.301310] XENBUS: Unable to read cpu state
> recent call last[ 86.307501] XENBUS: Unable to read cpu state
> ):
> File "/usr[ 86.313693] XENBUS: Unable to read cpu state
> /sbin/xend", lin[ 86.320070] XENBUS: Unable to read cpu state
> e 158, in <modul[ 86.326362] XENBUS: Unable to read cpu state
> e>
> sys.exit[ 86.332740] XENBUS: Unable to read cpu state
> (main())
> File[ 86.339068] XENBUS: Unable to read cpu state
> "/usr/sbin/xend[ 86.345378] XENBUS: Unable to read cpu state
> ", line 133, in [ 86.351797] xenbus_probe_devices backend
> main
> start_[ 86.358882] xenbus_probe_devices failed xenbus_directory
> [ 86.364819] backend_probe_and_watch devices probed ok
> blktapctrl()
> [ 86.370770] backend_probe_and_watch watch add ok ok
> [ 86.377463] backend_probe_and_watch all done
> [ 86.382232] xenbus_probe_devices device
> File "/usr/sbin/[ 86.386832] xenbus_probe_devices failed
> xenbus_directory
> [ 86.394018] frontend_probe_and_watch devices probed ok
> xend", line 108,[ 86.400078] frontend_probe_and_watch watch add ok ok
> [ 86.406862] frontend_probe_and_watch all done
> in start_blktapctrl
> start_daemon("blktapctrl", "")
> File "/usr/sbin/xend", line 74, in start_daemon
> os.execvp(daemon, (daemon,) + args)
> File "/usr/lib/python2.5/os.py", line 354, in execvp
> _execvpe(file, args)
> File "/usr/lib/python2.5/os.py", line 392, in _execvpe
> func(fullname, *argrest)
> OSError: [Errno 2] No such file or directory
> [3]root@intel_5500_server:/root>
> I got above error, I don't know what cause it.
> then I find I don't have /dev/xvda in Dom0,
> so I create it with
>
> mknod /dev/xvda b 202 0
> mknod /dev/xvda1 b 202 1
>
> in dom0, question: do I need to create it in dom0? if not, where it can be
> create?
>
> Then I create a image " dd if=/dev/zero of=domU.image bs=1M count=1000"
>
>
> and my domU cfg file is
>
> disk = ['file:/domU.image,xvda1,w']
> kernel = "/vmlinuz-2.6.31.13"
> vcpus=2
> extra = "3 selinux=0 enforcing=0"
> root = "/dev/xvda1 ro"
>
> (if I don't create /dev/xvda1 in dom0, it will tell me no /dev/xvda1
> device)
> does this file miss something?
>
> I don't use initrd file for dom0 or domU, does it need for me?
>
> then I boot domU with
> xm create -f xx.cfg -c,
> I got
>
> [ 0.266829] List of all partitions:
> [ 0.266841] ca01 1024000 xvda1 driver: vbd
> [ 0.266849] No filesystem could mount root, tried: ext3 vfat msdos
> iso9660
> [ 0.266868] Kernel panic - not syncing: VFS: Unable to mount root fs on
> unknown-block(202,1)
> [ 0.266879] Pid: 1, comm: swapper Not tainted 2.6.31.13 #1
>
> any idea?
>
The image file is empty, so it doesn't contain a filesystem that could be
mounted in the guest.
You need to actually *INSTALL* a guest to the image file.
-- Pasi
> Thanks
> Lei
>
>
>
>
>
>
> On Sat, May 15, 2010 at 2:35 AM, Pasi Kärkkäinen <[4]pasik@xxxxxx> wrote:
>
> On Sat, May 15, 2010 at 02:00:10AM +0800, lei yang wrote:
> >
> > I'm confused.. Are you trying to use the _same_ disk for the
> guest as
> > you use for dom0?
> > You can't do that.
> >
> >
> > my dom0 use /dev/sda1, can I use domU with second partition
> /dev/sda2(may
> > be it correspond xvda2?)
> >
>
> Yes you can.
>
> Or you can use a file in dom0 filesystem. Or you can create LVM
> volumegroup
> and create an LVM volume for the guest.
> > if I can't use the second partition, hown do I use "an image file"
> , how
> > to produce a image file, and can I put it in /dev/sda1 to use
> >
>
> You can create an empty image file with "dd", for example, or with any
> another
> normal unix/linux tool.
>
> Usually the tool you use to install the guest OS can make guest image
> files for you.
> > Actually, I don't know what does mean by "disk =
> ['phy:/dev/sda,xvda,w' ]"
> > can you explan it for me
> >
>
> It means use xen "phy:" driver for the guest virtual disk, which
> requires you to
> specify a 'physical' block device in dom0. The block device in question
> is "/dev/sda",
> and it's mapped to be "/dev/xvda" in the guest. "w" means it's in
> read/write mode.
>
> Another options is to use "file:" driver, which enables you to use image
> files instead of block devices.
> > does virtual disk use the phy disk space in fact?
> >
>
> Depends how you set it up.
>
> Also I recommend you to start the guest with "xm create -f
> /etc/xen/<guest> -c",
> which opens up the guest terminal immediately and allows you to see the
> guest kernel boot process with all the messages.
>
> -- Pasi
> > Thanks
> > Lei
> >
> > You need to have another disk for the guest, or an image file, or
> lvm
> > volume..
> > > 4. I create the device by
> > > mknod /dev/xvda b 202 0
> > > mknod /dev/xvda1 b 202 1
> > > when I mount -t ext3 /dev/xvda /
> > >
> > > [5]root@intel_5500_server:/> mount -t ext3 /dev/xvda /
> > > mount: /dev/xvda is not a valid block device
> > > does it mean my xvda device driver is not loaded
> > >
> >
> > xvda is the device name in the *guest*, which means the guest
> kernel
> > will see the virtual disk as /dev/xvda.
> > -- Pasi
> >
> > --
> > "We learn from failure, not from success!"
> >
> > References
> >
> > Visible links
> > 1. mailto:[5]pasik@xxxxxx
>
> --
> "We learn from failure, not from success!"
>
> References
>
> Visible links
> 1. mailto:root@intel_5500_server/root
> 2. mailto:root@intel_5500_server/root
> 3. mailto:root@intel_5500_server/root
> 4. mailto:pasik@xxxxxx
> 5. mailto:pasik@xxxxxx
_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users
|
Home