VPP does not support multiple interfaces in the same subnet.
Your scenario will be a configuration error once:
https://gerrit.fd.io/r/#/c/8057/
is committed.
/neale
From: <vpp-dev@lists.fd.io> on behalf of bindiya Kurle <bindiyaku...@gmail.com>
Date: Monday, 7 May 2018 at 07:27
To: "vpp-dev@lists.fd.io" <vpp-dev@lists.fd.io>
Subject: [vpp-dev] Query on VPP behaviour when IP from same subnet configured
on plain and vlan interface
Hi,
13.0.0.200 -------------------------------|GigabitEthernet1/0/0 (plain
interface)13.0.0.2
|
GigabitEthernet1/0/0.111(vlan interface)13.0.0.5 packet to send out
destination IP (13.0.0.200)
Fig 1.
I am trying to configure two IP’s belonging to same subnet on plain and a VLAN
interface(refer fig 1).While sending a packet, the ip4-lookup node is fetching
the dpoi_index pertaining to the VLAN interface which in-turn gives the
software index to VLAN interface in lookup.
If I try same scenario on Linux ,ping to the same destination IP(IP:
13.0.0.200) works as kernel pick up the plain interface route since that is the
1st route in its routing table.
FIB table entry:
13.0.0.2/32<http://13.0.0.2/32> pmtu: 0
unicast-ip4-chain
[@0]: dpo-load-balance: [proto:ip4 index:12 buckets:1 uRPF:16 to:[0:0]]
[0] [@2]: dpo-receive: 13.0.0.2 on GigabitEthernet1/0/0
13.0.0.5/32<http://13.0.0.5/32> pmtu: 0
unicast-ip4-chain
[@0]: dpo-load-balance: [proto:ip4 index:17 buckets:1 uRPF:22 to:[0:0]]
[0] [@2]: dpo-receive: 13.0.0.5 on GigabitEthernet1/0/0.111
13.0.0.200/32<http://13.0.0.200/32> pmtu: 0
UNRESOLVED
Questions :
1. Is there any specific reason why VPP always returns last entry added for
that prefix instead of 1st entry? Can Vpp behaviour be made similar to Linux
kernel Behaviour?
Regards,
Bindiya
