-
Notifications
You must be signed in to change notification settings - Fork 129
Description
I am trying to make some initial configuration through RESTful plugin. My vpp version is 22.10 and it has 2 DPDK interfaces TGE1 and TGE2. I'm using the linux-cp plugin of VPP. I hope to create a loopback interface, but the ligato seems not support linux-cp plugin configuration, so I write the lcp create command in the bootstrap.vpp file, and in my vpp.conf, I use exec bootstrap.vpp to execute the commands when vpp starting.
My bootstrap.vpp is:
create loopback interface
lcp create loop0 host-if loop0
lcp create TGE1 host-if eth1
lcp create TGE2 host-if eth2
Then I write a init.yaml and PUT it to the ligato RESTful plugin to configure the IP addresses and the states of the interfaces.
vppConfig:
interfaces:
- name: "loop0"
type: SOFTWARE_LOOPBACK
enabled: true
ipAddresses:
- 10.200.0.1/24
- fec0:2020::1/64
- name: "TGE1"
type: DPDK
enabled: true
ipAddresses:
- 10.200.20.10/24
- fec0:2023::1/96
- name: "TGE2"
type: DPDK
enabled: true
ipAddresses:
- 10.200.30.10/24
- fec0:2024::2/96However, I found that the ip addresses of TGE1 and TGE2 were configured correctly, but loop0 wasn't configured. Actually, when I use vppctl show int addr, I found a new loopback loop1 was created and configured with the ip addresses of loop0. The loop0 and the host-interface loop0 created by linux-cp stayed down and unconfigured.
I tried to solve the problem. I dumped the interfaces with the RESTful api, and found that the name of loop0 is actually UNTAGGED-loop0. I modified the YAML configuration above but it didn't work. The behavior was almost same: a new loopback interface was created and configured, but the original loop0 wasn't.
Is this a BUG of ligato or my YAML configuration is wrong? How can I configure a existed loopback interface rather than create a new one?