NIC Tags

NIC tags serve for mapping virtual networks to physical network interfaces. Danube Cloud comes with the admin NIC tag preconfigured, which is used for the admin network.

For other use cases, it is possible to add any any NIC tags. The following NIC tags can be created during installation:

  • external
  • internal
  • storage

After manually configuring NIC tags directly on the compute node, the compute node’s system information must be refreshed in the Danube Cloud web management. The new NIC tags will be present in the NIC tag dropdown menu when creating new network.


NIC tags are managed manually on compute nodes and it is the administrator’s job to make sure that different nodes don’t have NIC tags with the same name but different type. Removing one of the NIC tags with different type or setting the same type on NIC tag on all nodes will solve this issue.


external, internal, storage NIC tags are preconfigured in the Danube Cloud install script. However, it is up to the data center administrator to properly configure these additional NIC tags on all compute nodes.


After installing Danube Cloud or during the installation, it is recommended to configure the external NIC tag. (Even if it points to the same physical network interface/link aggregation as the admin NIC tag. If new network interfaces are added to the compute node in the future, it will be easier to just modify the NIC tag mapping and not the configuration of every virtual network).

Managing NIC Tags

The nictagadm command is used for working with NIC tags on the compute node. It is possible to add, delete or modify NIC tags with it. It also updates the /usbkey/config file, which makes the NIC tag configuration persistent over reboots.

There are 4 types of NIC tags available:

  • normal
  • aggregation
  • etherstub
  • overlay rule


Normal, aggregation, and etherstub NIC tags should only be managed using nictagadm; if you use dladm to create and/or delete these NIC tags, unexpected behavior might occur.


Overlay rule NIC tags are special because they are created based on the overlay_rules.json file. Once the first VM that uses an overlay is created, the overlay spawns into life.

See also

Overlay rule NIC tags are related to VXLAN configuration in virtual networks.

Adding normal/aggr NIC Tag

A NIC tag can be mapped to a physical interface by specifying the MAC address of the interface:

[root@node01 ~] nictagadm list
NAME           MACADDRESS         LINK           TYPE
admin          78:24:af:9c:3b:53  rge0           normal
[root@node01 ~] nictagadm add external 78:24:af:9c:3b:53
[root@node01 ~] nictagadm list
NAME           MACADDRESS         LINK           TYPE
external       78:24:af:9c:3b:53  rge0           normal
admin          78:24:af:9c:3b:53  rge0           normal

For mapping a NIC tag to a link aggregation interface, it is necessary to use the physical link aggregation name (e.g. aggr0) instead of the MAC address:

[root@node01 ~] nictagadm add external aggr0
[root@node01 ~] nictagadm list
NAME           MACADDRESS         LINK
external       -                  aggr0
admin          -                  aggr0

Adding etherstub NIC tag

Etherstubs are created with the command shown below. Please make sure that etherstub name ends with a number, otherwise you will end up with an error.

[root@node01 ~] nictagadm add -l stub0
[root@node01 ~] nictagadm list
NAME           MACADDRESS         LINK           TYPE
stub0          -                  -              etherstub

Adding overlay NIC tag

As opposed to other NIC tag types, overlays cannot be currently managed through the nictagadm command. Overlays are created based on overlay rules. Overlay rules are either created by manually editing the /var/run/smartdc/networking/overlay_rules.json file or by adding overlay_rule_<name> entries into the /usbkey/config file:

Example overlay rule configuration in /usbkey/config.
overlay_rule_my_overlay="-e vxlan -p vxlan/listen_ip=,vxlan/listen_port=4790 -s files -p files/config=/opt/custom/networking/my_overlay.json -p mtu=1400"
overlay_rule_foobar="-e vxlan -p vxlan/listen_ip=,vxlan/listen_port=4791 -s files -p files/config=/opt/custom/networking/other_overlay.json -p mtu=1400"
The configuration above will automatically create /var/run/smartdc/networking/overlay_rules.json after reboot.
[root@node01 ~] cat /var/run/smartdc/networking/overlay_rules.json
    "my_overlay": "-e vxlan -p vxlan/listen_ip=,vxlan/listen_port=4790 -s files -p files/config=/opt/custom/networking/my_overlay.json -p mtu=1400",
    "foobar": "-e vxlan -p vxlan/listen_ip=,vxlan/listen_port=4791 -s files -p files/config=/opt/custom/networking/other_overlay.json -p mtu=1400"

Deleting a NIC Tag

The nictagadm delete command should be used to delete NIC tags.

[root@node01 ~] nictagadm delete external
[root@node01 ~] nictagadm list
NAME           MACADDRESS         LINK           TYPE
admin          78:24:af:9c:3b:53  rge0           normal

Overlays must be deleted using both dladm command and removed manually from /var/run/smartdc/networking/overlay_rules.json. Additionally the overlay_rule_<name> entries should be removed from the persistent configuration in /usbkey/config.

[root@node01 ~] dladm show-overlay
my_overlay2233    mtu                rw   -   1400       1400      576-8900
my_overlay2233    vnetid             rw   -   2234       --        --
my_overlay2233    encap              r-   -   vxlan      --        vxlan
my_overlay2233    varpd/id           r-   -   1          --        --
my_overlay2233    vxlan/listen_ip    rw   y    --        --
my_overlay2233    vxlan/listen_port  rw   y   4700       4700      1-65535
my_overlay2233    search             r-   -   files      --        direct,
my_overlay2233    files/config       rw   y   /opt/custom/networking/my_overlay.json -- --

[root@node01 ~] dladm delete-overlay my_overlay2234
[root@node01 ~] nictagadm list
NAME           MACADDRESS         LINK           TYPE
admin          78:24:af:9c:3b:53  rge0           normal