[SOLVED] crash/reboot at VM starting

D

Deni74

Well-Known Member
Apr 29, 2013
67
1
48
Hello.
New server:
Supermicro MB X10DDW-iN
2xE5-2630. 128 Gb RAM. AOM 3108 HW Raid. 8 Segate Const., 4 SSD Intel (now off ).
All bios, firmware is updated. Bios: Integr. fake raid - Off, virt. tech. - On.
Proxmox 4.1 last version - no installation problems.

Crash / reboot at start any VM:
syslog [----] 71 L:[12464+25 12489/14574] *(1231708/1479893b) 0101 0x065 [*][X]
Mar 22 08:49:42 kvm pvedaemon[1453]: Use of inherited AUTOLOAD for non-method PVE::API2::Tasks::raise_param_exc() is deprecated at /usr/share/perl5/PVE/
Mar 22 08:49:43 kvm pvedaemon[1453]: Use of inherited AUTOLOAD for non-method PVE::API2::Tasks::raise_param_exc() is deprecated at /usr/share/perl5/PVE/
Mar 22 08:49:44 kvm pvedaemon[1453]: Use of inherited AUTOLOAD for non-method PVE::API2::Tasks::raise_param_exc() is deprecated at /usr/share/perl5/PVE/
Mar 22 08:49:45 kvm systemd-timesyncd[902]: interval/delta/delay/jitter/drift 64s/+0.001s/0.017s/0.000s/+0ppm
Mar 22 08:49:45 kvm pvedaemon[1452]: Use of inherited AUTOLOAD for non-method PVE::API2::Tasks::raise_param_exc() is deprecated at /usr/share/perl5/PVE/
Mar 22 08:49:46 kvm pvedaemon[1453]: <root@pam> starting task UPID:kvm:0000062B:00001B24:56F0DCFA:qmstart:103:root@pam:
Mar 22 08:49:46 kvm pvedaemon[1579]: start VM 103: UPID:kvm:0000062B:00001B24:56F0DCFA:qmstart:103:root@pam:
Mar 22 08:49:46 kvm pvedaemon[1453]: Use of inherited AUTOLOAD for non-method PVE::API2::Tasks::raise_param_exc() is deprecated at /usr/share/perl5/PVE/
Mar 22 08:49:47 kvm systemd[1]: Failed to reset devices.list on /system.slice: Invalid argument
Mar 22 08:49:47 kvm kernel: [ 70.677040] device tap103i0 entered promiscuous mode
Mar 22 08:49:47 kvm kernel: [ 70.683465] vmbr0: port 2(tap103i0) entered forwarding state
Mar 22 08:49:47 kvm kernel: [ 70.683475] vmbr0: port 2(tap103i0) entered forwarding state
Mar 22 08:49:47 kvm pvedaemon[1451]: No balloon device has been activated
Mar 22 08:49:47 kvm pvedaemon[1451]: Use of inherited AUTOLOAD for non-method PVE::API2::Tasks::raise_param_exc() is deprecated at /usr/share/perl5/PVE/
Mar 22 08:49:47 kvm pvedaemon[1451]: No balloon device has been activated
Mar 22 08:49:48 kvm pvedaemon[1451]: Use of inherited AUTOLOAD for non-method PVE::API2::Tasks::raise_param_exc() is deprecated at /usr/share/perl5/PVE/
Mar 22 08:49:48 kvm pvedaemon[1453]: No balloon device has been activated
Mar 22 08:49:49 kvm pvedaemon[1453]: No balloon device has been activated
Mar 22 08:49:49 kvm pvedaemon[1453]: Use of inherited AUTOLOAD for non-method PVE::API2::Tasks::raise_param_exc() is deprecated at /usr/share/perl5/PVE/
Mar 22 08:49:50 kvm pvedaemon[1451]: No balloon device has been activated
Mar 22 08:49:50 kvm kernel: [ 73.500362] kvm: zapping shadow pages for mmio generation wraparound
Mar 22 08:49:50 kvm kernel: [ 73.504534] kvm: zapping shadow pages for mmio generation wraparound
Mar 22 08:49:50 kvm pvedaemon[1451]: No balloon device has been activated
Mar 22 08:49:50 kvm pvedaemon[1451]: Use of inherited AUTOLOAD for non-method PVE::API2::Tasks::raise_param_exc() is deprecated at /usr/share/perl5/PVE/
Mar 22 08:49:51 kvm pvedaemon[1451]: No balloon device has been activated
Mar 22 08:49:51 kvm pvedaemon[1451]: No balloon device has been activated
Mar 22 08:49:51 kvm pvedaemon[1451]: Use of inherited AUTOLOAD for non-method PVE::API2::Tasks::raise_param_exc() is deprecated at /usr/share/perl5/PVE/
Mar 22 08:49:52 kvm pvedaemon[1451]: No balloon device has been activated
Mar 22 08:49:52 kvm pvedaemon[1451]: No balloon device has been activated
Mar 22 08:49:53 kvm pvedaemon[1451]: Use of inherited AUTOLOAD for non-method PVE::API2::Tasks::raise_param_exc() is deprecated at /usr/share/perl5/PVE/
Mar 22 08:49:53 kvm pvedaemon[1453]: No balloon device has been activated
Mar 22 08:49:53 kvm pvedaemon[1453]: No balloon device has been activated
Mar 22 08:49:54 kvm pvedaemon[1453]: Use of inherited AUTOLOAD for non-method PVE::API2::Tasks::raise_param_exc() is deprecated at /usr/share/perl5/PVE/
Mar 22 08:49:54 kvm pvedaemon[1453]: No balloon device has been activated
Mar 22 08:49:54 kvm pvedaemon[1453]: No balloon device has been activated
^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@( crash here)
Mar 22 08:52:19 kvm systemd-modules-load[475]: Module 'fuse' is builtin
Mar 22 08:52:19 kvm systemd-modules-load[475]: Inserted module 'vhost_net'
Mar 22 08:52:19 kvm hdparm[499]: Setting parameters of disc: (none).

I tried to change the various BIOS settings , but without success.
Any ideas?

Denis.
 
Last edited:
2xProcessor: Intel(R) Xeon(R) CPU E5-2630 v3 @ 2.40GHz
Speed: 2400 MHz
Core: 8
Core Active: 8
Manufacturer: Intel

/var/log/mcelog [----] 27 L:[ 1+16 17/ 35] *(566 /1091b) 0067 0x043
mcelog: Unsupported new Family 6 Model 3f CPU: only decoding architectural errors
mcelog: failed to prefill DIMM database from DMI data
mcelog: Unsupported new Family 6 Model 3f CPU: only decoding architectural errors
Hardware event. This is not a software error.
MCE 0
CPU 8 BANK 0 TSC 5d027e0b174.
RIP !INEXACT! 10:ffffffffa011f2b0
TIME 1458663441 Wed Mar 23 04:17:21 2016
MCG status:RIPV MCIP.
MCi status:
Error overflow
Uncorrected error
Error enabled
Processor context corrupt
MCA: Unknown Error 5
STATUS f200000000090005 MCGSTATUS 5
MCGCAP 7000c16 APICID 10 SOCKETID 1.
CPUID Vendor Intel Family 6 Model 63
mcelog: Unsupported new Family 6 Model 3f CPU: only decoding architectural errors
Hardware event. This is not a software error.
MCE 0
CPU 24 BANK 0 TSC 5d027e0b188.
RIP !INEXACT! 10:ffffffff812f4c09
TIME 1458663441 Wed Mar 23 04:17:21 2016
MCG status:RIPV MCIP.
MCi status:
Error overflow
Uncorrected error
Error enabled
Processor context corrupt
MCA: Unknown Error 5
STATUS f200000000090005 MCGSTATUS 5
MCGCAP 7000c16 APICID 11 SOCKETID 1.
CPUID Vendor Intel Family 6 Model 63

Now try mcelog with Proxmox 4.1
 
Last edited:
Already run quick memtest showed no errors.
mcelog Proxmox 4.1 is the same.
I'll try full memtest.
 
Last edited:
LnxBil said:
Shows the integrated management server errors?
Click to expand...

IPMI, Bios, Megaraid logs - only information messages like boot complete etc, and some warnings like power loss.
maybe it's bios settings? Now show settings. New MB - is not yet fully understood bios.
 
My newest supermicro server here is more than 5 years old, so it hopefully got fixed already, but I had problems when I enabled iommu in the bios. Disabling it solved every stability problem I ever had with a supermicro.
 
One thing about your bios settings: I had stability problems on HP hardware concerning energy efficiency. The machines crashed when I did not enforce non-energy-saving. So setting to maximum power everywhere. Worth a try on your machine.
 
PCI PERR/SERR Support could lead to other diagnosis also worth a try.

On older Xeon-based machines, I had problems with p-states, but your machine is very new so it should not suffer from that old problem.
 
My old Supermicro server 5 years old too - Supermicro X8DT3-F , i don't had problems (iommu - yes remember).
Energy efficiency now off, but.. I'll try find another bios settings like this. I think so - problem in energy efficiency.
 
I tried a lot of settings - same mestake. I don't know what can I do more. May be try install Proxmox on last version Debian?
 
If Debian Jessie works fine (I'm talking about the original Debian kernel), then the problem lies in the ubuntu longern support kernel proxmox is using and maybe there are some known bugs.

Do you encounter the bug also without starting a VM? It would be good to be able to reproduce the problem.

You can also install kdump-tools and configure your kernel commandline with crashkernel=256MB in order to get a real crash dump and investigate further. Unfortunately, this will get more and more time consuming :-/
 
  • Like
Reactions: Deni74
:( Proxmox installation on Debian 8.3 (Optional: Install Proxmox VE on Debian 8 Jessie (64 bit)) works not fine - the same.

Do you encounter the bug also without starting a VM?
Click to expand...
without starting - no problems.
kdump-tools - it's a sadly o_O :-/

Now trying original Debian KVM.
 
Try Debian KVM with libvirt and virual manager - same error... yes - bios settings or memory error
 
You must log in or register to reply here.
Top Bottom