Windows XP Multicore BSOD

[SebM]

Hello everybody,

I`ve got a problem with Windows XP VM on Proxmox.

During work on the VM it randomly crashes and shows very often different BSODs. (i.e. 0x1000000a)

Reconfiguring the VM from 2 Cores/Socket to 1 Cores/Socket did solve this problem. But Im not satisfied with this solution. Maybe you could give me a hint how to fix this issue?

Here are some information from the mainpage:

Version (package/version/build) pve-manager/1.5/4660
Kernel Version Linux 2.6.32-1-pve #1 SMP Fri Jan 15 11:37:39 CET 2010

Thanks for your help.

Greets
SebM

 

[tom]

post the output of 'pveversion -v', your are not using the latest version.

(I suggest you move to pvetest repository to test the latest kvm (we also plan to upload a new 2.6.32 this week))

 

[SebM]

Heres the output of pveversion -v:

pve-manager: 1.5-7 (pve-manager/1.5/4660)
running kernel: 2.6.32-1-pve
proxmox-ve-2.6.32: 1.5-4
pve-kernel-2.6.32-1-pve: 2.6.32-4
pve-kernel-2.6.24-10-pve: 2.6.24-21
pve-kernel-2.6.18-2-pve: 2.6.18-5
pve-kernel-2.6.18-1-pve: 2.6.18-4
qemu-server: 1.1-11
pve-firmware: 1.0-3
libpve-storage-perl: 1.0-10
vncterm: 0.9-2
vzctl: 3.0.23-1pve8
vzdump: 1.2-5
vzprocps: 2.0.11-1dso2
vzquota: 3.0.11-1
pve-qemu-kvm: 0.11.1-2
ksm-control-daemon: 1.0-3

But Ive got a newer host system that has this this problem too. Here the pveversion of that hostsystem:

pve-manager: 1.5-10 (pve-manager/1.5/4822)
running kernel: 2.6.32-2-pve
proxmox-ve-2.6.32: 1.5-7
pve-kernel-2.6.32-2-pve: 2.6.32-7
pve-kernel-2.6.18-2-pve: 2.6.18-5
qemu-server: 1.1-16
pve-firmware: 1.0-5
libpve-storage-perl: 1.0-13
vncterm: 0.9-2
vzctl: 3.0.23-1pve11
vzdump: 1.2-5
vzprocps: 2.0.11-1dso2
vzquota: 3.0.11-1
pve-qemu-kvm: 0.12.4-1
ksm-control-daemon: 1.0-3

These machines are working productive so I dont know if its that good to use the latest pvetest kvm.

But thanks so far.

Greets
SebM

 

[dietmar]

Just uploaded an update to the 'pvetest' repository - please can you test that?

 

[SebM]

So you want me to test the 'pvetest' repository? Is there a possible fix included?

Thank you.

 

[tom]

there is the latest KVM - and yes, pls test the 2.6.32 Kernel.

 

[gunforce]

Hi,
I have the same problem and pvetest didn't fix the error.
I only have it on a server with an 2 Xeon(R) CPU X5560. on my test server with i7 CPU 920 there no BSOD with the same Virtual Maschine.

best regards


proxmox-srv:~# pveversion -v
pve-manager: 1.6-9 (pve-manager/1.6/5307)
running kernel: 2.6.32-4-pve
proxmox-ve-2.6.32: 1.6-27
pve-kernel-2.6.32-4-pve: 2.6.32-27
pve-kernel-2.6.32-1-pve: 2.6.32-4
pve-kernel-2.6.35-1-pve: 2.6.35-7
qemu-server: 1.1-25
pve-firmware: 1.0-9
libpve-storage-perl: 1.0-16
vncterm: 0.9-2
vzctl: 3.0.24-1pve4
vzdump: 1.2-9
vzprocps: 2.0.11-1dso2
vzquota: 3.0.11-1
pve-qemu-kvm: 0.13.0-2
ksm-control-daemon: 1.0-4

 

[tom]

post the config from your windows guest, see /etc/qemu-server/VMID.conf

 

[gunforce]

Here is the config: 114.conf

ostype: wxp
memory: 1024
sockets: 1
name: VS2008_2
vlan0: e1000=F2:817:2B:07:53
scsi0: local:114/vm-106-disk-1.qcow2
localtime: 1
boot: dc
freeze: 0
cpuunits: 1000
acpi: 1
kvm: 1
onboot: 0
cores: 2
bootdisk: scsi0
scsi1: local:114/vm-106-disk-2.qcow2
vga: cirrus



Thanks

 

[tom]

any special reason why you use scsi for disks? the general rule for windows is IDE and raw format - best tested and widely used.

 

[gunforce]

for scsi there was no special reason, but for qcow2 there i can make snapshots. now i changed the vm to ide.

 

[whinpo]

I got the same kind of troubles with a Win2k8 x64...
I've tried in VIRTIO and IDE : same error 0x0000000a
My storage is on iSCSI

but didn't try with 1 Core only...
(I should maybe open another post as far as it's on Win2k8?)

neptune2:~# pveversion -v
pve-manager: 1.5-10 (pve-manager/1.5/4822)
running kernel: 2.6.32-2-pve
proxmox-ve-2.6.32: 1.5-8
pve-kernel-2.6.32-2-pve: 2.6.32-8
pve-kernel-2.6.24-10-pve: 2.6.24-21
qemu-server: 1.1-16
pve-firmware: 1.0-5
libpve-storage-perl: 1.0-13
vncterm: 0.9-2
vzctl: 3.0.24-1pve2
vzdump: 1.2-7
vzprocps: 2.0.11-1dso2
vzquota: 3.0.11-1
pve-qemu-kvm: 0.12.5-1
ksm-control-daemon: 1.0-4

name: Xenapp1
ide2: NFSMirror:iso/XenApp5_XA501_WS08_EN.iso,media=cdrom
vlan0: virtio=2A:44:AF:C6:DC:13
bootdisk: virtio0
ostype: w2k8
memory: 4096
sockets: 1
onboot: 0
description: 
cores: 2
boot: cad
freeze: 0
cpuunits: 1000
acpi: 1
kvm: 1
ide0: ProxmoxTest-SR:vm-104-disk-1
ide1: ProxmoxTest-SR:vm-104-disk-2

 

[tom]

use latest version from pvetest, better?

 

[whinpo]

tried also (from pvetest on Monday) => same problem
tried on a non-hyperthreaded machine also => idem

win32k.sys or ntkrlnmp.exe are the cause of the cras (says windbg)

 

[gunforce]

Hi
with ide i have the same problem
regards

 

[gunforce]

Hi All,
i thing i found the problem.

I disabeled some Bios settings:
Intel SpeedStep
Intel C-State
C1E Support
Hardware Prefetcher
Adjacent Cache Line Prefetch
Execute-Disable Bit Capability

Now I enabeled the following setting and wait if the error came back:
C1E Support
Hardware Prefetcher
Adjacent Cache Line Prefetch

A nice side effect is the my average load fall from 0,35 to 0,15.

Regards

 

[SebM]

Hello gunforce,

thanks for your answer.

So by disabling:

Intel SpeedStep
Intel C-State
Execute-Disable Bit Capability

The problem is fixed?

Greets

 

[gunforce]

disabling the following fixed the problem.

Intel SpeedStep
Intel C-State
C1E Support
Hardware Prefetcher
Adjacent Cache Line Prefetch
Execute-Disable Bit Capability

I had 4 days with no error. now i try to enable the settings, but i need some days to see if the error came back.

Regards

 

[tesquenure]

Hi Gunforce,

What about your test now ?

Regards,
Tes

 

[gunforce]

Hi All,

1. Test
C1E Support > enabled
Hardware Prefetcher > enabled
Adjacent Cache Line Prefetch > enabled
Intel C-State > disabled
2 Day Withou Error but slow again average load 0.35

2. Test
Hardware Prefetcher > enabled
Adjacent Cache Line Prefetch > enabled
C1E Support > disabled
Intel C-State > disabled
average load was back to 0.10
4 Days without errors

3. Test
Hardware Prefetcher > enabled
Adjacent Cache Line Prefetch > enabled
C1E Support > disabled
Intel SpeedStep with TurboMode > enabled
Intel C-State > disabled
2 Days without errors

Now i stop testing.

Result:
C1E Support > disabled
Intel C-State > disabled

Regards