CoreOS does NOT configure network form configdrive network_data.json

[stamak]

We have on OpenStack network with disabled dhcp (static IP)
When we spin up a CoreOS VM we do not have IP configures as CoreOS
does not take it from config drive
When we start Ubuntu or CentOS VM with cloud-init we have everything in place and
networking working as intended.

More detail below

core@node-0 ~ $ sudo mkdir /config_drive_mount
core@node-0 ~ $ sudo mount -o ro -t auto /dev/disk/by-label/config-2 /config_drive_mount
core@node-0 ~ $ sudo ls -la /config_drive_mount
total 24
drwxr-xr-x.  4 root root 16384 Jan  1  1970 .
drwxr-xr-x. 19 root root  4096 Dec  6 09:06 ..
drwxr-xr-x.  4 root root  2048 Dec  6 08:50 ec2
drwxr-xr-x. 10 root root  2048 Dec  6 08:50 openstack
core@node-0 ~ $ sudo ls -la /config_drive_mount/openstack/latest/
total 12
drwxr-xr-x.  2 root root 2048 Dec  6 08:50 .
drwxr-xr-x. 10 root root 2048 Dec  6 08:50 ..
-rwxr-xr-x.  1 root root 1918 Dec  6 08:50 meta_data.json
-rwxr-xr-x.  1 root root  970 Dec  6 08:50 network_data.json
-rwxr-xr-x.  1 root root    2 Dec  6 08:50 vendor_data.json
-rwxr-xr-x.  1 root root   14 Dec  6 08:50 vendor_data2.json
core@node-0 ~ $ sudo cat /config_drive_mount/openstack/latest/network_data.json
{
   "services":[
      {
         "type":"dns",
         "address":"8.8.4.4"
      },
      {
         "type":"dns",
         "address":"8.8.8.8"
      }
   ],
   "networks":[
      {
         "network_id":"a41735c0-4c4e-4b1a-a19e-874ce376bef6",
         "type":"ipv4",
         "services":[
            {
               "type":"dns",
               "address":"8.8.4.4"
            },
            {
               "type":"dns",
               "address":"8.8.8.8"
            }
         ],
         "netmask":"255.255.255.192",
         "link":"tap9bed847b-30",
         "routes":[
            {
               "netmask":"0.0.0.0",
               "network":"0.0.0.0",
               "gateway":"209.X.X.1"
            }
         ],
         "ip_address":"209.X.X.24",
         "id":"network0"
      },
   ],
   "links":[
      {
         "ethernet_mac_address":"fa:16:3e:f3:d5:8e",
         "mtu":1500,
         "type":"ovs",
         "id":"tap9bed847b-30",
         "vif_id":"9bed847b-3044-4b17-a046-30693c144a5b"
      },
   ]
}

Version

cat /etc/os-release
NAME="Container Linux by CoreOS"
ID=coreos
VERSION=2303.3.0
VERSION_ID=2303.3.0
BUILD_ID=2019-12-02-2049
PRETTY_NAME="Container Linux by CoreOS 2303.3.0 (Rhyolite)"
ANSI_COLOR="38;5;75"
HOME_URL="https://coreos.com/"
BUG_REPORT_URL="https://issues.coreos.com"
COREOS_BOARD="amd64-usr

[lucab]

Thanks for the report.

A few things on this:

• from your os-release, you seem to be using Container Linux. The bug tracker for that is at https://github.com/coreos/bugs, not here.
• Afterburn is the component in charge of sourcing instance metadata. Indeed at this point it does not support openstack-configdrive.
• our recommendation would be to use DHCP (for dynamic network configuration) and the metadata endpoint (for node introspection), which works more reliably and across cloud providers.

[lucab]

Closing in favor or coreos/bugs#2645.