It seems that there is some sort of race condition related to kvm on one of my servers where starting kvm virtual machines will cause networking to stop. while sleep 5 ; do ping -c4 8.8.8.8 && killall -1 qemu-system-x86_64; done To solve this I run the above shell code while starting kvms, if something goes wrong then it kills them all and networking works again. This first happened in March, so I think it was some Debian update in that time period that introduced the issue. The problem is that I don't reboot KVM servers very often so it's a little difficult to narrow the problem down. I don't know for sure that my other KVM servers don't have the issue, maybe they have just got lucky every time I booted them. Any suggestions for what to investigate? PS that's why the LUV server has had some downtime today. -- My Main Blog http://etbe.coker.com.au/ My Documents Blog http://doc.coker.com.au/