Debian 12 (from marketplace) netplan does not apply context with netplan

Community Support
1

I have 2 VMs, Debian 11 and Debian 12, both from OpenNebula Public Marketplace.
Both VMs have 2 NICs (public and private bridges which are not the problem here).

I have configured public IPs on both VMs.

The context generated /etc/network/interfaces in Debian 11 works fine, with all routes working. But the context generated /etc/netplan/50-cloud-init.yaml does not work and I end with a system with no default route.

I have done nothing, just selected the virtual networks to instance both VMs. I can’t understand wtf is going on with netplan in OpenNebula’s Debian 12 image (an why tf they use netplan).

Can you help me?

image
image862×643 89.5 KB

image

2

Checking /var/log/syslog, if I have only public NIC/Network: no ip route at all.

Checking /var/log/syslog, if I have both public and private NIC/Network: Error: Conflicting default route declarations for IPv4 (table: main, metric: 0), first declared in eth1 but also in eth0.

Is there any known issue in one-context netplan config? There are no errors when I’m using /etc/network/interfaces in Debian 11, for example, and I guess it’s not a netplan issue.

3

Here is the /etc/netplan/50-one-context.yaml generated by one-context:

# Generated by one-context
network:
  version: 2
  renderer: networkd
  ethernets:
    eth0:
      addresses:
        - <public-ip>/32
      routes:
        - to: "0.0.0.0/0"
          via: <public-gateway>
          metric: 0
    eth1:
      addresses:
        - 10.10.114.160/16
      routes:
        - to: "0.0.0.0/0"
          via: 10.10.114.1
          metric: 0

Here is the default route after the “Error: Conflicting default route declarations for IPv4 (table: main, metric: 0), first declared in eth1 but also in eth0”:

default via 10.10.114.1 dev eth1 proto static 
10.10.0.0/16 dev eth1 proto kernel scope link src 10.10.114.160

I can’t get the ip route from my public NIC, even if I start the VM only with public NIC.

4

I changed NETCFG_TYPE following rollback instructions found in the one-apps releases page to get it working in Debian 12 cloud image.

NETCFG_TYPE="interfaces"

Even though this solved the route problem, the problem between one-context and netplan remains.

5

Can you send the VM instance template for this debian 12 issue to see if we can reproduce it ?

1 Like
6

The image used is Debian 12 from OpenNebula Marketplace (Apps section). Now it’s working fine using old /etc/network/interfaces and ifupdown.

Template (excluding NETCFG_TYPE = “interfaces” line):

AUTOMATIC_DS_REQUIREMENTS = “("CLUSTERS/ID" @> 100)”
AUTOMATIC_NIC_REQUIREMENTS = “("CLUSTERS/ID" @> 100)”
AUTOMATIC_REQUIREMENTS = “(CLUSTER_ID = 100) & !(PUBLIC_CLOUD = YES) & !(PIN_POLICY = PINNED)”
CONTEXT = [
DISK_ID = “1”,
ETH0_DNS = “8.8.8.8 8.8.4.4”,
ETH0_EXTERNAL = “”,
ETH0_GATEWAY = “”,
ETH0_IP = “”,
ETH0_IP6 = “”,
ETH0_IP6_GATEWAY = “”,
ETH0_IP6_METHOD = “”,
ETH0_IP6_METRIC = “”,
ETH0_IP6_PREFIX_LENGTH = “”,
ETH0_IP6_ULA = “”,
ETH0_MAC = “02:00:00:e9:a6:c6”,
ETH0_MASK = “255.255.255.255”,
ETH0_METHOD = “static”,
ETH0_METRIC = “”,
ETH0_MTU = “”,
ETH0_NETWORK = “”,
ETH0_SEARCH_DOMAIN = “”,
ETH0_VLAN_ID = “”,
ETH0_VROUTER_IP = “”,
ETH0_VROUTER_IP6 = “”,
ETH0_VROUTER_MANAGEMENT = “”,
ETH1_DNS = “8.8.8.8 8.8.4.4”,
ETH1_EXTERNAL = “”,
ETH1_GATEWAY = “10.10.114.1”,
ETH1_IP = “10.10.114.160”,
ETH1_IP6 = “”,
ETH1_IP6_GATEWAY = “”,
ETH1_IP6_METHOD = “”,
ETH1_IP6_METRIC = “”,
ETH1_IP6_PREFIX_LENGTH = “”,
ETH1_IP6_ULA = “”,
ETH1_MAC = “02:00:0a:0a:72:a0”,
ETH1_MASK = “255.255.0.0”,
ETH1_METHOD = “”,
ETH1_METRIC = “”,
ETH1_MTU = “”,
ETH1_NETWORK = “10.10.0.0”,
ETH1_SEARCH_DOMAIN = “”,
ETH1_VLAN_ID = “”,
ETH1_VROUTER_IP = “”,
ETH1_VROUTER_IP6 = “”,
ETH1_VROUTER_MANAGEMENT = “”,
NETWORK = “YES”,
SSH_PUBLIC_KEY = “”,
TARGET = “hda” ]
CPU = “0.3”
DISK = [
ALLOW_ORPHANS = “FORMAT”,
CLONE = “YES”,
CLONE_TARGET = “SYSTEM”,
CLUSTER_ID = “100”,
DATASTORE = “Default-D02”,
DATASTORE_ID = “101”,
DEV_PREFIX = “vd”,
DISK_ID = “0”,
DISK_SNAPSHOT_TOTAL_SIZE = “0”,
DISK_TYPE = “FILE”,
DRIVER = “qcow2”,
FORMAT = “qcow2”,
IMAGE = “Debian 12 D02 Image”,
IMAGE_ID = “24”,
IMAGE_STATE = “2”,
LN_TARGET = “SYSTEM”,
ORIGINAL_SIZE = “20480”,
READONLY = “NO”,
SAVE = “NO”,
SIZE = “20480”,
SOURCE = “/var/lib/one//datastores/101/c8c3121cc1996024293cb50ee39a4cd1”,
TARGET = “vda”,
TM_MAD = “shared”,
TM_MAD_SYSTEM = “ssh”,
TYPE = “FILE” ]
GRAPHICS = [
LISTEN = “0.0.0.0”,
PORT = “5949”,
TYPE = “VNC” ]
MEMORY = “2048”
MEMORY_MAX = “16384”
MEMORY_RESIZE_MODE = “BALLOONING”
NIC = [
AR_ID = “0”,
BRIDGE = “br0”,
BRIDGE_TYPE = “linux”,
CLUSTER_ID = “100”,
GATEWAY = “”,
IP = “”,
MAC = “02:00:00:e9:a6:c6”,
METHOD = “static”,
NAME = “NIC0”,
NETWORK = “Ponte externa D02”,
NETWORK_ID = “32”,
NIC_ID = “0”,
SECURITY_GROUPS = “0”,
TARGET = “one-49-0”,
VN_MAD = “bridge” ]
NIC = [
AR_ID = “1”,
BRIDGE = “br1”,
BRIDGE_TYPE = “linux”,
CLUSTER_ID = “100”,
GATEWAY = “10.10.114.1”,
IP = “10.10.114.160”,
MAC = “02:00:0a:0a:72:a0”,
NAME = “NIC1”,
NETWORK = “Ponte interna D02”,
NETWORK_ID = “31”,
NETWORK_UNAME = “leandro”,
NIC_ID = “1”,
SECURITY_GROUPS = “0”,
TARGET = “one-49-1”,
VN_MAD = “bridge” ]
OS = [
UUID = “3f811896-1e3b-4094-87f6-6532415e29da” ]
SECURITY_GROUP_RULE = [
PROTOCOL = “ALL”,
RULE_TYPE = “OUTBOUND”,
SECURITY_GROUP_ID = “0”,
SECURITY_GROUP_NAME = “default” ]
SECURITY_GROUP_RULE = [
PROTOCOL = “ALL”,
RULE_TYPE = “INBOUND”,
SECURITY_GROUP_ID = “0”,
SECURITY_GROUP_NAME = “default” ]
TEMPLATE_ID = “23”
TM_MAD_SYSTEM = “ssh”
VCPU = “1”
VCPU_MAX = “16”
VMID = “49”

7

Try issuing different metric values for each of the interfaces configuration. For example set ETH0_METRIC to 100 and ETH1_METRIC to 200. In the yaml you sent, both are 0, which could be the cause of the conflict.

Alternatively, use only 1 default gw and static routes.