Bug#524571: Bug triggered by core/VMs

[Laurent Mahieux <debian@xenpad.eu>]

  Hi,

  I read this thread while looking for the cause of unexplained dom0
reboots (unattended reboot on a headless server, so no idea what
happened).

  The crashes seem to be triggered by two factors for me:
  - heavy I/O (backup of dom0 by rsync)
  - more VMs than cores (including dom0)

  On a bi-xeon, linux sees 4 cores (thanks to hyperthreading). As long
as I'm running 3 domU or less everything works (that's
3 domU + 1 dom0 <= 4 cores)
  If I start one more domU, no matter how small or inactive, on the
next backup, the server will crash and reboot 50% of the time.

  Hope this observation can help narrow it down.

  Best regards,
   Laurent