Why VM cannot get the public IP I have from provider?

Question

I have a KVM-Host (on Debian) with

root@kvm-host:/# cat /etc/network/interfaces

auto  eth0
iface eth0 inet manual

auto vmbr0
iface vmbr0 inet static
  address 176.x.y.133
  netmask 255.255.255.224
  gateway 176.x.y.129 # Provider's GW IP
  bridge_ports eth0
  bridge_stp off
  bridge_fd 0


root@kvm-host:/# route -n

Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
176.x.y.128     0.0.0.0         255.255.255.224 U     0      0        0 vmbr0
0.0.0.0         176.x.y.129     0.0.0.0         UG    0      0        0 vmbr0


root@kvm-host:/# arp -an

? (176.x.y.155) at <incomplete> on vmbr0
? (176.x.y.129) at xx:yy:zz:11:22:33 [ether] on vmbr0


root@kvm-host:/# ifconfig

1: lo: ...
2: eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state UP qlen 1000
    link/ether AA:BB:CC:DD:EE:FF brd ff:ff:ff:ff:ff:ff
    inet6 fe80::a60:6eff:feDD:EEFF/64 scope link 
       valid_lft forever preferred_lft forever
3: vmbr0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc noqueue state UNKNOWN 
    link/ether AA:BB:CC:DD:EE:FF brd ff:ff:ff:ff:ff:ff
    inet 176.x.y.133/27 brd 176.x.y.159 scope global vmbr0
    inet6 fe80::a60:6eff:feDD:EEFF/64 scope link 
       valid_lft forever preferred_lft forever
4: venet0: <BROADCAST,POINTOPOINT,NOARP,UP,LOWER_UP> mtu 1500 qdisc noqueue state UNKNOWN 
    link/void 
    inet6 fe80::1/128 scope link 
       valid_lft forever preferred_lft forever

I have public IPs: 176.x.y.133 (KVM-Host), 176.x.y.151, 176.x.y.155.

Now I'm trying to get the network on each VM:

user@VM:/$ cat /etc/network/interfaces

auto lo
iface lo inet loopback

auto eth0
iface eth0 inet static
  address 176.x.y.155
  netmask 255.255.255.255
  pointtopoint 176.x.y.133
  gateway 176.x.y.133

I uninstalled the network-manager.

Next steps:

user@VM:/$ sudo ifup eth0
RTNETLINK answers: File exists
Failed to bring up eth0


user@VM:/$ ping 176.x.y.133
connect: Network is unreachable

root@VM:/# route -n
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface

(i.e. empty, no even default gateway)

user@VM:/$ sudo add default via 176.x.y.133
RTNETLINK answers: no such process

user@VM:/$ sudo arp -a
(i.e. empty)

user@VM:/$ sudo arp -s 176.x.y.133 AA:BB:CC:DD:EE:FF
SIOCSARP Network is unreachable

I can't even set static routes (MAC-addr./IP-addr.).

What could be the problem to set a network on the VM? (I hope I provided all the needed information to be specific)

You have Hetzner or OVH? Contact the provider; they have strange network setups that require some weird setup. — Michael Hampton, Jul 16 '13 at 18:42
Hetzner :) Already written them, but they mean "virtualization is not supproted as a service". :( — static, Jul 16 '13 at 23:59
Actually supported, but you need to buy additional ip for each VM and allocate unique mac address in web panel. I'm using ESXi 5.1 with 5 public ip without any problem — ALex_hha, Jul 21 '13 at 22:00
it doesn't work for me: I need the Routed variant, not a bridged one (that's why you need to allocate unique mac address in robot-pannel) — static, Jul 21 '13 at 22:02
With Hetzner you'll either need to use proxy ARP on the host or manually set the MAC address to the one generated in their Robot. — James O'Gorman, Jul 25 '13 at 12:07

Rob Bos · Answer 1 · 2013-07-22T02:48:26.510

1

Check that the MAC address is not randomly generated; many ISPs will only allocate IPs to MAC addresses that are issued by valid providers. I was able to get an IP on a machine by taking a known-good one, incrementing the last octet by one.

Edit: make sure "brctl show vmbr0" shows something sensible. Something like

xen-br0         8000.90b11c316da4       no              eth0

edited Jul 22 '13 at 02:48

answered Jul 22 '13 at 02:37

Rob Bos

854
7
8

yep: bridge name bridge id STP enabled interfaces vmbr0 8000.xxxxxxxxxx no eth0 tap100i0 tap101i0 – static Jul 25 '13 at 17:51

score 0 · Answer 2 · answered Jul 27 '13 at 06:21

On HW node (host), /etc/network/interface could look like:

# This file describes the network interfaces available on your system
# and how to activate them. For more information, see interfaces(5).

# Which interface are to be started at boot
auto lo vmbr0

# The loopback network interface
iface lo inet loopback

# The bridge for VM    
iface vmbr0 inet static
        address 176.x.y.133
        netmask 255.255.255.224
        network 176.x.y.128
        broadcast 176.x.y.159
        gateway 176.x.y.129 # Provider's GW IP
        bridge_ports eth0
        bridge_maxwait 2
        up /sbin/ifconfig eth0 inet 0.0.0.0 promisc
        up /sbin/sysctl -w net.ipv4.conf.br0.proxy_arp=1
        up /sbin/sysctl -w net.ipv4.conf.default.forwarding=1
        up /sbin/sysctl -w net.ipv4.ip_forward=1

Of course, you may prefer to create a vm-routing file in /etc/sysctl.d/ to store sysctl config.

On the VM Host, you must be able to use same network configuration:

iface eth0 inet static
        address 176.x.y.155
        netmask 255.255.255.224
        network 176.x.y.128
        broadcast 176.x.y.159
        gateway 176.x.y.129 # Provider's GW IP

Why VM cannot get the public IP I have from provider?

2 Answers2

Linked