Re: [Xen-users] Xen 3.4.3 - randomly crash of Dom0 and DomU

[Jonathan Tripathy <jonnyt@xxxxxxxxxxx>]

On 15/06/10 07:52, Jonathan Tripathy wrote:

On 07/06/10 08:58, Felix Botner wrote:

Hi everyone,

from time to time (but at least one crash in 24h) the system crashes (neither 
Dom0 nor DomU is accessible via network or console) with no usefull 
information in any log file and i have no idea what the problem could be.
Log files (syslog, kern.log, messages, xend.log) on both systems, Dom0 and 
DomU, show nothing, no kernel oops or anything.

The setup is:

   * a Dom0 (64Bit) with a 2.6.32-xen, Debian Sid based kernel 
     (linux-source-2.6.32 2.6.32-13) with xen-3.4.3 (build with 
     the standard options)

   * and one paravirtualized Debian Etch based linux DomU (32Bit) with a
     Debian Etch based kernel 2.6.18-xen (also tried other Debian Lenny based 
     DomU Kernels, 2.6.26-xen and 2.6.30-xen)

   * DomU's network interface is connected via bridge

   DomU cgf:
     name = 'name'
     kernel = "/opt/ucs2.2_xen/vmlinuz-2.6.18-ucs153-xen-686"
     memory = 2500
     vcpus = 1
     disk = ['file:/var/lib/xen/name/disk,xvda,w' ]
     vif = [ 'mac=00:16:3e:8e:8b:08' ]
     boot = 'd'
     ramdisk = "/opt/ucs2.2_xen/initrd.img-2.6.18-ucs153-xen-686"
     root = "/dev/xvda1 ro"
     extra = 'xencons=tty console=hvc0'

   Dom0 /etc/xen/xend-config.sxp:
     (xen-api-server ((unix)))
     (xend-http-server no)
     (xend-unix-server no)
     (xend-unix-xmlrpc-server yes)
     (xend-tcp-xmlrpc-server no)
     (xend-relocation-server no)
     (xend-relocation-ssl-server no)
     (network-script network-bridge)
     (vif-script vif-bridge)
     (dom0-min-mem 196)
     (enable-dom0-ballooning yes)
     (dom0-cpus 0)
     (vncpasswd '')

   grub/menu.lst
     title           Xen 3.4.3 / UCS, kernel 2.6.32-ucs5-xen-amd64
     root            (hd0,0)
     kernel          /xen-3.4.3.gz dom0_mem=3072M
     module          /vmlinuz-2.6.32-ucs5-xen-amd64 
                       root=/dev/mapper/vg_ucs-rootfs ro root2fstype=ext3  
                       splash=silent video=vga16:off nosplash rootdelay=5
                       clock=pit loglevel=0 nopat console=tty0 fbcon=map:2
     module          /initrd.img-2.6.32-ucs5-xen-amd64


Is there any known problem running a 32Bit DomU in a 64 Bit Dom0?
Could anyone give me a hint, how to debug this situation or where i can find 
more information?

Feel free to ask for more information about the system or the setup.

Many thanks

  

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users

So I did a test to see if I could replicate what Fexlix was experiencing.

I left 3 DomUs running (2PV and 1 HVM) on my Ubuntu 10.04 Dom0 running overnight (with 2 of them pinging an internet server, the other one was a pfsense firewall running in HVM), and in the morning the server had indeed crashed (Had to hard reset it). The kernel I'm using is Jerem'y 2.6.32.15 pvops kernel. Both Dom0 and DomUs are 64-bit. I'm using Xen 3.4.3.

Xend.log says "Xend exited with status 0)" at around 4:12am in the morning.
The last thing that Xend-debug says is this (no time given):

1. Exception in thread HttpServer:
2. Traceback (most recent call last):
3.   File "/usr/lib/python2.6/threading.py", line 532, in __bootstrap_inner
4.     self.run()
5.   File "/usr/lib/python2.6/threading.py", line 484, in run
6.     self.__target(*self.__args, **self.__kwargs)
7.   File "/usr/lib/python2.6/dist-packages/xen/web/httpserver.py", line 312, in run
8.     (sock, addr) = self.accept()
9.   File "/usr/lib/python2.6/dist-packages/xen/web/httpserver.py", line 331, in accept
10.     return self.socket.accept()
11.   File "/usr/lib/python2.6/socket.py", line 195, in accept
12.     sock, addr = self._sock.accept()
13. error: [Errno 22] Invalid argument

Any help is appreciated

I should probably mention that for some reason when my system starts up, or when I try and start Xend manually, the init script launch 2 xend processes. Could this have caused the problem?

Thanks

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users

Ok so a little bit of an update.

My pfsense HVM machine has a PCI NIC passed through to it. I ran my test again, but this time the system crashed in front of me. But what actually happens, is that my Dom0 crashed (can't move mouse or keyboard) however its video still works. I can see the pings failing on the screen, which indicates that the pfsense HVM has also crashed. I have no way to restart the system except by power.

Seems to me like a PCI host controller fault??

Thanks

_______________________________________________
Xen-users mailing list
Xen-users@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-users