Bug#799253: virtio ens3 network interface

2015-09-20 Thread Martin Pitt
Hello Harald, Harald Dunkel [2015-09-18 7:38 +0200]: > I have still hosts that were continuously upgraded since Lenny, > now waiting to be migrated to Jessie. How long are you going to > support the old NIC naming? At least for stretch (obviously), in stretch+1 we'll most probably add a

Bug#799253: virtio ens3 network interface

2015-09-19 Thread Harald Dunkel
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On 09/18/15 07:38, Harald Dunkel wrote: > Hi Martin, > > On 09/18/15 04:36, Martin Pitt wrote: > >> Ah, I see. With the current naming system you have to disable the >> predictable names differerently now, either by booting with net.ifnames=0 or

Bug#799253: virtio ens3 network interface

2015-09-17 Thread Harald Dunkel
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Package: udev Version: 226-2 Since I upgraded my KVM host the NIC (virtio) is called "ens3" instead of "eth0". Looking at udev's changelog for 226-2 I had the impression that this is not supposed to happen. Regards Harri -BEGIN PGP

Bug#799253: virtio ens3 network interface

2015-09-17 Thread Martin Pitt
Control: tag -1 moreinfo Hello Harald, Harald Dunkel [2015-09-17 11:12 +0200]: > Since I upgraded my KVM host the NIC (virtio) is called "ens3" > instead of "eth0". Looking at udev's changelog for 226-2 I had > the impression that this is not supposed to happen. Indeed. Did the upgrade create a

Bug#799253: virtio ens3 network interface

2015-09-17 Thread Harald Dunkel
On 09/17/15 15:38, Martin Pitt wrote: > > Okay, so you do have such a device indeed. I don't see what's wrong > then. Can you please run > > sh -x /var/lib/dpkg/info/udev.postinst configure 220-6 > > as root and copy the entire output? > Ah, I see. The udev.postinst script is hardware

Bug#799253: virtio ens3 network interface

2015-09-17 Thread Martin Pitt
Hello Harald, Harald Dunkel [2015-09-17 14:59 +0200]: > root@usbpc:~# ls -lR /sys/bus/virtio/drivers/virtio_net > lrwxrwxrwx 1 root root0 Sep 17 14:42 virtio0 -> > ../../../../devices/pci:00/:00:03.0/virtio0 Okay, so you do have such a device indeed. I don't see what's wrong then.

Bug#799253: virtio ens3 network interface

2015-09-17 Thread Harald Dunkel
Hi Martin, here is more information: root@usbpc:~# ls -al /etc/systemd/network ls: cannot access /etc/systemd/network: No such file or directory root@usbpc:~# grep udev /var/log/aptitude [UPGRADE] libudev1:amd64 220-6 -> 226-2 [UPGRADE] udev:amd64 220-6 -> 226-2 root@usbpc:~# ls -lR

Bug#799253: virtio ens3 network interface

2015-09-17 Thread Martin Pitt
Hey Harald, Harald Dunkel [2015-09-17 20:02 +0200]: > > In such a scenario you can already not rely on stable interface names, as > > they would change with the changing hardware (unless you have an existing > > 70-persistent-net-names.rules from jessie, then it's okay again). > > > >

Bug#799253: virtio ens3 network interface

2015-09-17 Thread Harald Dunkel
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 Hi Martin, On 09/18/15 04:36, Martin Pitt wrote: > > Ah, I see. With the current naming system you have to disable the predictable > names differerently now, either by booting with net.ifnames=0 or by creating > an empty