2.6.24-7 does not start as xen domU

Bug #190870 reported by Jan Evert van Grootheest
4
Affects Status Importance Assigned to Milestone
linux (Ubuntu)
Invalid
Medium
Unassigned

Bug Description

Binary package hint: linux-image-2.6.24-7-xen

I decided to try to use this package (amd64 arch) for a domU, odo, on a xen server. For fun.
Odo, FYI, is the domU used to test with and play around. The other domU handles all the mail and nfs serving.

The server is debian stable (dom0, quark, is etch with kernel 2.6.18-5 and xen is 3.0.3) all amd64.

Unfortunately it doesn't even start properly. As far as I can tell this is caused by "XENBUS: Device with no driver: device/vbd/2065" just before the initrd scripts start.
The vbd/2065 is configured in the configuration file to show up as sdb1.
On dom0 this vbd is an lvm volume.

Probably you'd like to know that the odo userspace is also debian etch.
There was a hitch when installing this kernel due to module-init-tools (this package wants 3.3-pre11, I believe). I force-depends-version-ed dpkg when installing the kernel. So there is still the old module-init-tools (3.3-pre4 from etch).
I decided not to install that (yet) because it depends on a newer libc (unfortunately). Better to start small with disrupting the system (-:

I'm not sure what other info to provide. Feel free to ask.

Here's the console from the create:
quark:~# xm create -c odo.cfg
Using config file "/etc/xen/odo.cfg".
Started domain odo
[ 0.000000] Linux version 2.6.24-7-xen (buildd@crested) (gcc version 4.2.3 (Ubuntu 4.2.3-1ubuntu1)) #1 SMP Thu Feb 7 02:53:39 UTC 2008 (Ubuntu 2.6.24-4.6-generic)
[ 0.000000] Command line: root=/dev/sdb1 ro
[ 0.000000] BIOS-provided physical RAM map:
[ 0.000000] Xen: 0000000000000000 - 0000000010800000 (usable)
[ 0.000000] end_pfn_map = 67584
[2892231.778961] Zone PFN ranges:
[2892231.778963] DMA 0 -> 4096
[2892231.778965] DMA32 4096 -> 1048576
[2892231.778967] Normal 1048576 -> 1048576
[2892231.778969] Movable zone start PFN for each node
[2892231.778970] early_node_map[1] active PFN ranges
[2892231.778972] 0: 0 -> 67584
[2892231.789869] No mptable found.
[2892231.795008] PERCPU: Allocating 22368 bytes of per cpu data
[2892231.795041] Built 1 zonelists in Zone order, mobility grouping on. Total pages: 66660
[2892231.795046] Kernel command line: root=/dev/sdb1 ro
[2892231.795753] Initializing CPU#0
[2892231.795924] PID hash table entries: 2048 (order: 11, 16384 bytes)
[2892231.795964] Xen reported: 1800.063 MHz processor.
[ 0.031186] console [xvc0] enabled
[ 0.031230] Console: colour dummy device 80x25
[ 0.031608] Dentry cache hash table entries: 65536 (order: 7, 524288 bytes)
[ 0.032090] Inode-cache hash table entries: 32768 (order: 6, 262144 bytes)
[ 0.032166] Software IO TLB disabled
[ 0.035548] Memory: 233896k/270336k available (2467k kernel code, 27924k reserved, 1313k data, 216k init)
[ 0.119710] Calibrating delay using timer specific routine.. 4511.89 BogoMIPS (lpj=9023785)
[ 0.119777] Security Framework initialized
[ 0.119786] SELinux: Disabled at boot.
[ 0.119795] AppArmor: AppArmor initialized
[ 0.119800] Failure registering capabilities with primary security module.
[ 0.119824] Mount-cache hash table entries: 256
[ 0.119969] CPU: L1 I Cache: 64K (64 bytes/line), D cache 64K (64 bytes/line)
[ 0.119973] CPU: L2 Cache: 512K (64 bytes/line)
[ 0.119992] SMP alternatives: switching to UP code
[ 0.120587] Freeing SMP alternatives: 22k freed
[ 0.120758] Early unpacking initramfs... done
[ 0.143866] Brought up 1 CPUs
[ 0.144579] net_namespace: 120 bytes
[ 0.144592] failed to set up cpufreq notifier
[ 0.156712] Time: 165:165:165 Date: 165/165/65
[ 0.156763] NET: Registered protocol family 16
[ 0.157560] Brought up 1 CPUs
[ 0.157579] PCI: Fatal: No config space access function found
[ 0.157587] PCI: setting up Xen PCI frontend stub
[ 0.158149] ACPI: Interpreter disabled.
[ 0.158161] Linux Plug and Play Support v0.97 (c) Adam Belay
[ 0.158214] pnp: PnP ACPI: disabled
[ 0.158334] xen_mem: Initialising balloon driver.
[ 0.158850] Setting mem allocation to 262144 kiB
[ 0.159188] PCI: System does not support PCI
[ 0.159197] PCI: System does not support PCI
[ 0.166227] NET: Registered protocol family 8
[ 0.166236] NET: Registered protocol family 20
[ 0.166343] AppArmor: AppArmor Filesystem Enabled
[ 0.166801] NET: Registered protocol family 2
[ 0.169620] Time: xen clocksource has been installed.
[ 0.201576] IP route cache hash table entries: 4096 (order: 3, 32768 bytes)
[ 0.201810] TCP established hash table entries: 16384 (order: 6, 262144 bytes)
[ 0.201994] TCP bind hash table entries: 16384 (order: 6, 262144 bytes)
[ 0.202191] TCP: Hash tables configured (established 16384 bind 16384)
[ 0.202195] TCP reno registered
[ 0.213649] checking if image is initramfs... it is
[ 0.237192] Freeing initrd memory: 17500k freed
[ 0.253805] audit: initializing netlink socket (disabled)
[ 0.253828] audit(1202716419.689:1): initialized
[ 0.254030] VFS: Disk quotas dquot_6.5.1
[ 0.254055] Dquot-cache hash table entries: 512 (order 0, 4096 bytes)
[ 0.254146] io scheduler noop registered
[ 0.254150] io scheduler anticipatory registered
[ 0.254152] io scheduler deadline registered
[ 0.254160] io scheduler cfq registered (default)
[ 0.254392] Xen virtual console successfully installed as xvc0
[ 0.254451] Event-channel device installed.
[ 0.283605] rtc: IRQ 8 is not free.
[ 0.283688] Linux agpgart interface v0.102
[ 0.283732] Serial: 8250/16550 driver $Revision: 1.90 $ 4 ports, IRQ sharing enabled
[ 0.284683] RAMDISK driver initialized: 16 RAM disks of 65536K size 1024 blocksize
[ 0.284753] input: Macintosh mouse button emulation as /devices/virtual/input/input0
[ 0.284878] PNP: No PS/2 controller found. Probing ports directly.
[ 0.285895] i8042.c: No controller found.
[ 0.288875] mice: PS/2 mouse device common for all mice
[ 0.288923] cpuidle: using governor ladder
[ 0.289007] NET: Registered protocol family 1
[ 0.289064] registered taskstats version 1
[ 0.289073] XENBUS: Device with no driver: device/vbd/2065
[ 0.289076] XENBUS: Device with no driver: device/vif/0
[ 0.289090] Magic number: 1:252:3141
[ 0.289205] /build/buildd/linux-2.6.24/debian/build/custom-source-xen/drivers/rtc/hctosys.c: unable to open rtc device (rtc0)
[ 0.289217] Freeing unused kernel memory: 216k freed
Loading, please wait...
Begin: Loading essential drivers... ...
Done.
Begin: Running /scripts/init-premount ...
FATAL: Error inserting fan (/lib/modules/2.6.24-7-xen/kernel/drivers/acpi/fan.ko): No such device
WARNING: Error inserting processor (/lib/modules/2.6.24-7-xen/kernel/drivers/acpi/processor.ko): No such device
[ 0.569693] thermal: Unknown symbol acpi_processor_set_thermal_limit
FATAL: Error inserting thermal (/lib/modules/2.6.24-7-xen/kernel/drivers/acpi/thermal.ko): Unknown symbol in module, or unknown parameter (see dmesg)
Done.
Begin: Mounting root file system... ...
Begin: Running /scripts/local-top ...
[ 1.977353] Uniform Multi-Platform E-IDE driver Revision: 7.00alpha2
[ 1.977368] ide: Assuming 50MHz system bus speed for PIO modes; override with idebus=xx
[ 4.255472] device-mapper: uevent: version 1.0.3
[ 4.255520] device-mapper: ioctl: 4.12.0-ioctl (2007-10-02) initialised: <email address hidden>
Done.
Begin: Waiting for root file system... ...
Done.
        Check root= bootarg cat /proc/cmdline
        or missing modules, devices: cat /proc/modules ls /dev
ALERT! /dev/sdb1 does not exist. Dropping to a shell!

BusyBox v1.1.3 (Debian 1:1.1.3-4) Built-in shell (ash)
Enter 'help' for a list of built-in commands.

/bin/sh: can't access tty; job control turned off
(initramfs)

Revision history for this message
Jan Evert van Grootheest (j-e-van-grootheest) wrote :

When rummaging through sysfs, I noticed this:
(initramfs) ls -l /sys/bus/xen/devices/vbd-2065/
-r--r--r-- 1 0 0 4096 Feb 12 17:31 devtype
-r--r--r-- 1 0 0 4096 Feb 12 17:31 nodename
drwxr-xr-x 2 0 0 0 Feb 12 17:31 power
lrwxrwxrwx 1 0 0 0 Feb 12 17:31 subsystem -> ../../../bus/xen
-rw-r--r-- 1 0 0 4096 Feb 11 07:53 uevent

while a normal domU has this:
suzy2:~# ls -l /sys/bus/xen/devices/vbd-2053/
total 0
lrwxrwxrwx 1 root root 0 Feb 12 18:33 block:sda5 -> ../../../block/sda5
lrwxrwxrwx 1 root root 0 Feb 12 18:33 bus -> ../../../bus/xen
-r--r--r-- 1 root root 4096 Feb 12 18:33 devtype
lrwxrwxrwx 1 root root 0 Feb 12 18:33 driver -> ../../../bus/xen/drivers/vbd
-r--r--r-- 1 root root 4096 Feb 12 18:33 nodename
drwxr-xr-x 2 root root 0 Jan 8 21:34 power
lrwxrwxrwx 1 root root 0 Feb 12 18:33 subsystem -> ../../../bus/xen
--w------- 1 root root 4096 Feb 12 18:33 uevent

So the driver and block:xxx links are missing.

I noticed that CONFIG_SYSFS_DEPRECATED is not set in the config.
Might that have anything to do with this?

Revision history for this message
Jan Evert van Grootheest (j-e-van-grootheest) wrote :
Revision history for this message
Leann Ogasawara (leannogasawara) wrote :

You may want toalso test linux-image-2.6.24-10-xen just to verify this still exists. Thanks.

Changed in linux:
assignee: nobody → ubuntu-kernel-team
importance: Undecided → Medium
status: New → Triaged
Revision history for this message
Jan Evert van Grootheest (j-e-van-grootheest) wrote :

Also found out that the block backend in the kernel config (in -8) in Ubuntu is a module, while in debian it is build-in.
So I'll also be making an initrd that has blkback in it. This is, of course, very likely the show stopper.

Might make sense to set this to build-in simply because otherwise this package is useless?

Revision history for this message
Jan Evert van Grootheest (j-e-van-grootheest) wrote :

linux-image-2.6.24-10-xen same story.
Still need to get around adding the xen block backend to the initrd (dayjob needs time as well).

Revision history for this message
Lionel Porcheron (lionel.porcheron) wrote :

Workaround:
add to /etc/initramfs-tools/modules:
xenblk
blktap
blkbck

and launch update-initramfs -u

Thanks PierreF (http://launchpad.net/~pierre-fersing) for your help ;-).

But now, I'm (still) faced to #196485. If you can check how it behaves for you... Thanks

Revision history for this message
Jan Evert van Grootheest (j-e-van-grootheest) wrote :

The workaround described by Lionel works.
Only xenblk is needed to make a domU boot.

Next stop: getting the network interface operational...
It first does this:
Synthesizing the initial hotplug events...done.
Waiting for /dev to be fully populated...[ 17.336715] netfront: Initialising virtual ethernet driver.
done.

And then somewhat later:
Setting up networking....
Configuring network interfaces...Internet Software Consortium DHCP Client 2.0pl5
Copyright 1995, 1996, 1997, 1998, 1999 The Internet Software Consortium.
All rights reserved.

Please contribute if you find this software useful.
For info, please visit http://www.isc.org/dhcp-contrib.html

[ 20.297349] NET: Registered protocol family 17
Listening on LPF/eth0/00:11:09:84:95:20
Sending on LPF/eth0/00:11:09:84:95:20
Sending on Socket/fallback/fallback-net
DHCPREQUEST on eth0 to 255.255.255.255 port 67
DHCPREQUEST on eth0 to 255.255.255.255 port 67
DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 3
DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 5
DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 11
DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 10
DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 19
DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 7
DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 6
No DHCPOFFERS received.
Trying recorded lease 192.168.1.114
PING 192.168.1.11 (192.168.1.11) 56(84) bytes of data.

--- 192.168.1.11 ping statistics ---
1 packets transmitted, 0 received, +1 errors, 100% packet loss, time 0ms

No working leases in persistent database.

Exiting.

Failed to bring up eth0.
done.

Revision history for this message
Jan Evert van Grootheest (j-e-van-grootheest) wrote :

This is the matching interface in dom0:
vif12.0 Link encap:Ethernet HWaddr FE:FF:FF:FF:FF:FF
          inet6 addr: fe80::fcff:ffff:feff:ffff/64 Scope:Link
          UP BROADCAST RUNNING NOARP MTU:1500 Metric:1
          RX packets:5 errors:0 dropped:0 overruns:0 frame:0
          TX packets:68 errors:0 dropped:27 overruns:0 carrier:0
          collisions:0 txqueuelen:0
          RX bytes:1640 (1.6 KiB) TX bytes:8565 (8.3 KiB)

I wonder why those 27 tx frames were dropped...
Also, the rx packets were successfully received by the dhcp server. At least some of the tx packets are responses.

Revision history for this message
Jan Evert van Grootheest (j-e-van-grootheest) wrote :

2.6.24-12 still has the same network problem. It appears to drop all received packets on the floor.
Without any explanation.

Revision history for this message
Maurício Magalhães (mauriciommagalhaes) wrote :

Hi all!

linux-image-2.6.24-12-xen same story.

sudo xm create linux.cfg -c

[...]

[ 0.263416] audit: initializing netlink socket (disabled)
[ 0.263430] audit(1205842523.020:1): initialized
[ 0.263583] VFS: Disk quotas dquot_6.5.1
[ 0.263613] Dquot-cache hash table entries: 1024 (order 0, 4096 bytes)
[ 0.263707] io scheduler noop registered
[ 0.263711] io scheduler anticipatory registered
[ 0.263714] io scheduler deadline registered
[ 0.263729] io scheduler cfq registered (default)
[ 0.263948] Xen virtual console successfully installed as xvc0
[ 0.263986] Event-channel device installed.
[ 0.298738] rtc: IRQ 8 is not free.
[ 0.299377] RAMDISK driver initialized: 16 RAM disks of 65536K size 1024 blocksize
[ 0.299447] input: Macintosh mouse button emulation as /devices/virtual/input/input0
[ 0.299569] PNP: No PS/2 controller found. Probing ports directly.
[ 0.300385] i8042.c: No controller found.
[ 0.303681] mice: PS/2 mouse device common for all mice
[ 0.303728] cpuidle: using governor ladder
[ 0.303817] NET: Registered protocol family 1
[ 0.303847] Using IPI No-Shortcut mode
[ 0.303876] registered taskstats version 1
[ 0.303886] XENBUS: Device with no driver: device/vbd/2049
[ 0.303889] XENBUS: Device with no driver: device/vbd/2050
[ 0.303893] XENBUS: Device with no driver: device/vif/0
[ 0.303896] XENBUS: Device with no driver: device/console/0
[ 0.303915] Magic number: 1:252:3141
[ 0.304164] Freeing unused kernel memory: 208k freed
Loading, please wait...
Begin: Loading essential drivers... ...
[ 0.693139] fuse init (API version 7.9)
[ 0.721353] thermal: Unknown symbol acpi_processor_set_thermal_limit
Done.
Begin: Running /scripts/init-premount ...
Done.
Begin: Mounting root file system... ...
Begin: Running /scripts/local-top ...
Done.
Begin: Waiting for root file system... ...

Done.
        Check root= bootarg cat /proc/cmdline
        or missing modules, devices: cat /proc/modules ls /dev
ALERT! /dev/sda2 does not exist. Dropping to a shell!

BusyBox v1.1.3 (Debian 1:1.1.3-5ubuntu12) Built-in shell (ash)
Enter 'help' for a list of built-in commands.

(initramfs)
########################################3
My linux.conf

#
# Configuration file for the Xen instance ecolx01, created
# by xen-tools 3.8 on Fri Mar 14 13:48:27 2008.
#
# Kernel + memory size
#
kernel = '/boot/vmlinuz-2.6.24-12-xen'
ramdisk = '/boot/initrd.img-2.6.24-12-xen'
memory = '128'
#
# Disk device(s).
#
root = '/dev/sda2 ro'
disk = [
                  'tap:aio:/home/xen/domains/ecolx01/swap.img,sda1,w',
                  'tap:aio:/home/xen/domains/ecolx01/disk.img,sda2,w',
              ]
#
# Hostname
#
name = 'ecolx01'
#
# Networking
#
vif = [ 'ip=10.32.76.100,mac=00:16:3E:C1:4B:A9' ]
#vif = [ 'bridge=xenbr0']
#vif = [ 'bridge=xenbr0' ]
#
# Behaviour
#
on_poweroff = 'destroy'
on_reboot = 'restart'
on_crash = 'restart'

Revision history for this message
Maurício Magalhães (mauriciommagalhaes) wrote :

Ignore my previous post, because it worked

Workaround:
add to /etc/initramfs-tools/modules:
xenblk
blktap
blkbck

and launch update-initramfs -u

Tanks

Revision history for this message
Jan Evert van Grootheest (j-e-van-grootheest) wrote : Re: [Bug 190870] Re: 2.6.24-7 does not start as xen domU

Maurício Magalhães wrote:
>
> Ignore my previous post, because it worked
>
> Workaround:
> add to /etc/initramfs-tools/modules:
> xenblk
> blktap
> blkbck
>
> and launch update-initramfs -u
>
> Tanks
>
>
Actually you only need xenblk in a domU.

I am interested whether you get the network operational. I can't get it
to receive anything.
And I'm stuck there.

Thanks,
Jan Evert

Revision history for this message
Jan Evert van Grootheest (j-e-van-grootheest) wrote :

So it was not really a bug, just that the block access is now as a module instead of built-in.
So I'm setting it to invalid (that's about the best I can guess).

Also in 2.6.24.-16 it is back to built-in.

And the networking is better handled in #204010.

Changed in linux:
status: Triaged → Invalid
Revision history for this message
Launchpad Janitor (janitor) wrote : Kernel team bugs

Per a decision made by the Ubuntu Kernel Team, bugs will longer be assigned to the ubuntu-kernel-team in Launchpad as part of the bug triage process. The ubuntu-kernel-team is being unassigned from this bug report. Refer to https://wiki.ubuntu.com/KernelTeamBugPolicies for more information. Thanks.

To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.