VIP segments, like 10.0.1.16/28, does no longer work

[uablrek]

After merge of https://github.com/Nordix/Meridio/pull/178 the VIP addresses are assigned to the nsc interface instead of lo.

/ # ip addr show dev nsc
38: nsc@if37: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 16000 qdisc noqueue state UP group default 
    link/ether 42:ae:eb:e2:ea:21 brd ff:ff:ff:ff:ff:ff link-netnsid 0
    inet 172.16.2.6/24 brd 172.16.2.255 scope global nsc
       valid_lft forever preferred_lft forever
    inet 10.0.0.1/32 scope global nsc
       valid_lft forever preferred_lft forever
    inet 10.0.0.16/28 brd 10.0.0.31 scope global nsc
       valid_lft forever preferred_lft forever
    inet6 1000::1:a00:10/124 scope global nodad 
       valid_lft forever preferred_lft forever
    inet6 1000::1:a00:1/128 scope global nodad 
       valid_lft forever preferred_lft forever
    inet6 fd00::ac10:206/120 scope global nodad 
       valid_lft forever preferred_lft forever
    inet6 fe80::40ae:ebff:fee2:ea21/64 scope link 
       valid_lft forever preferred_lft forever

This means that for instance 10.0.0.16/28 is no longer a segment (as on lo) but one single address 10.0.0.16 with a /28 mask.

[LionelJouin]

The plan is to break the prefixes and add every IPs to the interface.

[uablrek]

Are NSM cool with that?

[LionelJouin]

For now yes, but there might be new features to add prefixes on the loopback interface. Not sure when this will happen.

[uablrek]

I actually think that specifying a vip segment will not be used to configure say 4, or even 255 addresses but something like; 1000::/64.

[LionelJouin]

I agree, we could have some limit. I will check again if VIP range is needed.