Lately, I have been working to configure a centos65 server to enable VLAN tagging, this server is going to act as VM Host (kvm). On the network-switch level, the ports were configured to enable tagging, with following setup
vlan_a -> native vlan_b -> tagged
The server has a single interface -> eth0. (configured in trunk mode)
So while experimenting with VLAN tagging on this server, I had the following setup
eth0 - pointing to native vlan (10.20.12.12) eth0.43 -pointing to tagged vlan (10.21.8.4)
I would restart network services to enable this configuration. The problem I faced with this setup was, the tagged interface would be pingable and reachable, but the interface pointing to native vlan, was not reachable. Upon further inspection of routing table, I noticed the following
Destination Gateway Genmask Flags MSS Window irtt Iface 10.21.8.4 0.0.0.0 255.255.255.0 U 0 0 0 eth0.43 10.20.12.12 0.0.0.0 255.255.254.0 U 0 0 0 eth0 169.254.0.0 0.0.0.0 255.255.0.0 U 0 0 0 eth0 169.254.0.0 0.0.0.0 255.255.0.0 U 0 0 0 eth0.43 0.0.0.0 10.21.8.1 0.0.0.0 UG 0 0 0 eth0.43
So here is my question
In a tagged VLAN setup (for a single interface) for Linux systems, is it default that only the tagged interface comes up with default gateway? What is the best way to ensure that untagged interface is accessible as well?
To further elaborate, I did add the default gateway for untagged interface (eth0) and there on it was pingable and accessible. But is that ok as a practice to add default gateways for multiple interfaces in this kind of setup?
I have read about alternatives like openvswitch and network namespaces.
What are the specific advantages I have in using alternatives like openvswitch and network namespaces vs traditionally configuring linux system to handle this