Bug#569990: qemu-kvm: VMs don't start anymore

2010-03-12 Thread Michael Tokarev
Toni Mueller wrote at Tue, 16 Feb 2010 11:46:04 +0100:
> Hi,
> 
> On Mon, 15.02.2010 at 18:36:19 +0100, Toni Mueller  wrote:
>> I'm running into a severe problem, where I'm unsure about the origin.
>> Please feel free to re-assign as appropriate.
> 
>> three of them running Etch, Lenny, and Sid. It currently looks like the
>> Etch and Lenny VMs come up nicely, but the one with Sid hangs with this
>> error:
>>
>> Clocksource tsc unstable (telta = large negative number)

This message is to be expected and it is not harmful at all, kernel
can cope with tsc instability just fine, at least since 2.6.22 or so.

Speaking of Etch, the instability is a bit more important, but the
kernel still can do the right thing (and there should be no such
message printed with Etch because kernel does not have such message
if I remember correctly).

Try booting with different clock sources in guests (with Etch it is
not possible unfortunately, but Lenny+ is ok).  It is clocksource=XXX
on the kernel command line, or /sys/devices/system/clocksource/clocksource0
(can be changed at runtime).  For kvm the best clocksource is kvm-clock
(if the kernel recognizes it anyway, modern kernels does).

And while at it, again, please try the current version, which is
0.12.3+dfsg-4.  Or grab the .deb file from my site.

> I've now created a brand-new virtual machine, installing Unstable, but
> keeping the Lenny kernel. It turns out that the same machine starts ok
> with the Lenny kernel, but hangs with this error using the Sid kernel.

It's only kernel problem.  Ok, kvm+kernel, and since kernel usually
works without kvm it must be a kvm issue.  But clock is something
which is difficult to deal with in this situation.  For one, try
disabling any cpufreq things you may have on the host, and since
you're using Athlon X2, which inherently has unstable tsc (really
unsynced, between cores), the only more-or-less reliable way to
run a kvm guest on such machine is to pin it to a single core.
Or to use kvm-clock, which should work better.

> My current guess is that these components (Sid kernel, qemu-kvm) don't
> play well together, and, given the breakage I see with at least
> OpenBSD, I'm inclined to assume that the problem originates with
> qemu-kvm.
> 
> NB: On the virtual machine, I have the -i486 variants of the respective
> kernels installed. On the virtual machine where I originally reported
> the error for, I had the -i686 kernel variants installed.

Is there any difference between i686 and i484 variants?

Thanks!

/mjt



-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org



Bug#569990: qemu-kvm: VMs don't start anymore

2010-03-12 Thread Michael Tokarev
Toni Mueller wrote at Tue, 16 Feb 2010 16:58:13 +0100:
> Hi,
[]
> on a different machine, running Testing/i386, the VM boots nicely to
> the login prompt using -i686 variant of the 2.6.32 kernel, but there is
> no network. On the host system, I can see "tap0", and on tap0, I can
> see the DHCP requests from the guest, but the packets don't make it to
> the bridge interface which they are making it to on the Lenny/amd64
> machine. On the guest, when logged in via console, I can see eth0, but
> (naturally) no IP number on it. I've changed all encounters of the
> Realtek driver to specify e1000 instead, to no avail.

If you see on the tap device some packets coming from guest but no
response, check the host side - if the bridges are set up correctly,
if the firewall rules (or ebtables) forbids the guest's packets etc.
This is a problem not related to the $subject line, and lies within
your host settings, where kvm package does not try to be smart and
provides only over-simplistic, I'd say example-only, network script
(see /etc/kvm/kvm-ifup).  It is up to either the user or some higher-
level management tool (libvirt etc) to set up networking.  And this
has nothing to do with the guest NIC driver - if you see the problem
on host.

Thanks!

/mjt



-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org



Bug#569990: qemu-kvm: VMs don't start anymore

2010-02-16 Thread Toni Mueller

Hi,

one more data point:

On Tue, 16.02.2010 at 11:46:04 +0100, Toni Mueller  wrote:
> NB: On the virtual machine, I have the -i486 variants of the respective
> kernels installed. On the virtual machine where I originally reported
> the error for, I had the -i686 kernel variants installed.

on a different machine, running Testing/i386, the VM boots nicely to
the login prompt using -i686 variant of the 2.6.32 kernel, but there is
no network. On the host system, I can see "tap0", and on tap0, I can
see the DHCP requests from the guest, but the packets don't make it to
the bridge interface which they are making it to on the Lenny/amd64
machine. On the guest, when logged in via console, I can see eth0, but
(naturally) no IP number on it. I've changed all encounters of the
Realtek driver to specify e1000 instead, to no avail.

I'll dig into my options as to which kind of network I can have (will
check out virtio and vde2), and report my findings here.


Kind regards,
--Toni++



-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20100216155813.11695.qm...@oak.oeko.net



Bug#569990: qemu-kvm: VMs don't start anymore

2010-02-16 Thread Toni Mueller

Hi,

On Mon, 15.02.2010 at 18:36:19 +0100, Toni Mueller  wrote:
> I'm running into a severe problem, where I'm unsure about the origin.
> Please feel free to re-assign as appropriate.

> three of them running Etch, Lenny, and Sid. It currently looks like the
> Etch and Lenny VMs come up nicely, but the one with Sid hangs with this
> error:
> 
> Clocksource tsc unstable (telta = large negative number)

I've now created a brand-new virtual machine, installing Unstable, but
keeping the Lenny kernel. It turns out that the same machine starts ok
with the Lenny kernel, but hangs with this error using the Sid kernel.

My current guess is that these components (Sid kernel, qemu-kvm) don't
play well together, and, given the breakage I see with at least
OpenBSD, I'm inclined to assume that the problem originates with
qemu-kvm.

NB: On the virtual machine, I have the -i486 variants of the respective
kernels installed. On the virtual machine where I originally reported
the error for, I had the -i686 kernel variants installed.


Kind regards,
--Toni++




-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: http://lists.debian.org/20100216104604.19661.qm...@oak.oeko.net



Bug#569990: qemu-kvm: VMs don't start anymore

2010-02-15 Thread Toni Mueller
Package: qemu-kvm
Version: 0.11.0+dfsg-1~bpo50+1
Severity: important


Hi,

I'm running into a severe problem, where I'm unsure about the origin.
Please feel free to re-assign as appropriate.

I have various virtual machines, running various operating systems, and
three of them running Etch, Lenny, and Sid. It currently looks like the
Etch and Lenny VMs come up nicely, but the one with Sid hangs with this
error:

Clocksource tsc unstable (telta = large negative number)

The actual number varies, but the boot process hangs at this point for
at least 10 minutes (I've lost patience after that). The Lenny VM has
2.6.26 while the Sid VM has 2.6.30, 2.6.31, and 2.6.32 kernels (several
versions of these, actually).

This problem occurred "recently", but it now occurs with kvm (0.85bpo)
instead of qemu-kvm, too.

I'm starting my VMs with this command line (sample):

kvm -localtime -no-acpi -m 512 \
-net nic,model=rtl8139,macaddr=52:54:00:12:34:56,vlan=0 \
-net tap,vlan=0 /path/to/image/file/sid1.img &

which worked nicely for more than a year.

What gives?


Kind regards,
--Toni++


-- Package-specific info:


selected information from lshal(1):



/proc/cpuinfo:

processor   : 0
vendor_id   : AuthenticAMD
cpu family  : 15
model   : 107
model name  : AMD Athlon(tm) 64 X2 Dual Core Processor 4400+
stepping: 1
cpu MHz : 2300.000
cache size  : 512 KB
physical id : 0
siblings: 2
core id : 0
cpu cores   : 2
apicid  : 0
initial apicid  : 0
fpu : yes
fpu_exception   : yes
cpuid level : 1
wp  : yes
flags   : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov 
pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt rdtscp lm 
3dnowext 3dnow rep_good pni cx16 lahf_lm cmp_legacy svm extapic cr8_legacy 
3dnowprefetch
bogomips: 4625.17
TLB size: 1024 4K pages
clflush size: 64
cache_alignment : 64
address sizes   : 40 bits physical, 48 bits virtual
power management: ts fid vid ttp tm stc 100mhzsteps

processor   : 1
vendor_id   : AuthenticAMD
cpu family  : 15
model   : 107
model name  : AMD Athlon(tm) 64 X2 Dual Core Processor 4400+
stepping: 1
cpu MHz : 2300.000
cache size  : 512 KB
physical id : 0
siblings: 2
core id : 1
cpu cores   : 2
apicid  : 1
initial apicid  : 1
fpu : yes
fpu_exception   : yes
cpuid level : 1
wp  : yes
flags   : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov 
pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt rdtscp lm 
3dnowext 3dnow rep_good pni cx16 lahf_lm cmp_legacy svm extapic cr8_legacy 
3dnowprefetch
bogomips: 4625.17
TLB size: 1024 4K pages
clflush size: 64
cache_alignment : 64
address sizes   : 40 bits physical, 48 bits virtual
power management: ts fid vid ttp tm stc 100mhzsteps




-- System Information:
Debian Release: 5.0.4
  APT prefers stable
  APT policy: (990, 'stable'), (500, 'proposed-updates'), (450, 'testing'), 
(250, 'unstable'), (1, 'experimental')
Architecture: amd64 (x86_64)

Kernel: Linux 2.6.26-2-amd64 (SMP w/2 CPU cores)
Locale: LANG=de_DE.utf8, LC_CTYPE=en_US.utf8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/bash

Versions of packages qemu-kvm depends on:
ii  adduser   3.110  add and remove users and groups
ii  bridge-utils  1.4-5  Utilities for configuring the Linu
ii  iproute   20080725-2 networking and traffic control too
ii  libasound21.0.16-2   ALSA library
ii  libbluetooth2 3.36-1 Library to use the BlueZ Linux Blu
ii  libbrlapi0.5  3.10~r3724-1+lenny1braille display access via BRLTTY 
ii  libc6 2.7-18lenny2   GNU C Library: Shared libraries
ii  libcurl3-gnutls   7.18.2-8lenny3 Multi-protocol file transfer libra
ii  libgnutls26   2.4.2-6+lenny2 the GNU TLS library - runtime libr
ii  libncurses5   5.7+20081213-1 shared libraries for terminal hand
ii  libpci3   1:3.0.0-6  Linux PCI Utilities (shared librar
ii  libpulse0 0.9.10-3+lenny1PulseAudio client libraries
ii  libsasl2-22.1.22.dfsg1-23+lenny1 Cyrus SASL - authentication abstra
ii  libsdl1.2debian   1.2.13-2   Simple DirectMedia Layer
ii  libvdeplug2   2.2.2-3Virtual Distributed Ethernet - Plu
ii  libx11-6  2:1.1.5-2  X11 client-side library
ii  python2.5.2-3An interactive high-level object-o
ii  zlib1g1:1.2.3.3.dfsg-12  compression library - runtime

Versions of packages qemu-kvm recommends:
ii  linux-image-2.6.26-2-amd 2.6.26-21lenny3 Linux 2.6.26 image on AMD64
ii  linux-image-2.6.26-2-xen 2.6.26-21lenny3 Linux 2.6.26 image on AMD64, oldst

Versions of packages qemu-kvm suggests:
ii  d