Capif error when running register-and-onboard command

[Charles_Bailly]

Goal:
Trying to setup Capif with the IMM NetApp.

Result:
Errors when running manually the prepare.sh script:

requests.exceptions.ConnectionError: HTTPConnectionPool(host=‘capifcore’, port=8080): Max retries exceeded with url: /register (Caused by NewConnectionError(’<urllib3.connection.HTTPConnection object at 0x00000252E3118E80>: Failed to establish a new connection: [Errno 11001] getaddrinfo failed’))

Screenshot of the error:

Steps to reproduce:

1. Run the Capif services (sudo ./run.sh, got the “All Capif services are running” message)

2. Run the NEF emulator. Emulator starts successfully, but unclear if the unboarding with Capif went well or not.
   

   

3. Run the ./prepare.sh command. I’m doing it manually since I wanted to test Capif locally for now (=> not using a container for our NetApp).

Host environment:
Windows 10
Git bash
Docker-for-Windows
NEF emulator 1.6.1
Recent CAPIF version (from last week, develop branch)

/etc/host file: tried with localhost, then localhost and 172.17.0.1:

File structure:

Content of capif_onboarding folder:

Capif_registration json file:

Is there something I missed in the procedure to setup Capif?
Thank you in advance for your help.

[sxarism]

Hi Charles.
Do you run/deploy CAPIF-NEF-netapp on the same host?

[sxarism]

Also, the prepare.sh should be run inside the netapp container not the host.

[sxarism]

Have a look if you can on the dummy netapp repo

GitHub - EVOLVED-5G/dummy-netapp

[Charles_Bailly]

Hi Stavros,

Thank you for the quick answer.

Yes, Capif and Nef are running on the same host. I tried to reproduce what I saw on the dummy netapp and the videos from Ricardo and Dimitris.

However, I may have missunderstood the part about the prepare.sh script. Is is required only when using a container? Should I only use a container when working with Capif? Usually I’m working locally first (running the NetApp through my IDE) because it’s more convenient, then I try with a containerized version when its ready.

[sxarism]

Concerning the IP 172… etc, I think is visible only from containers on the same network.
So I would propose to work from the containerized version.

[Charles_Bailly]

Ok, I will focus on the containerized version.

One more question then: how can I check within the container that the capif registration and onboarding was successfull? So far, I’m getting the following error:

FileNotFoundError: [Errno 2] No such file or directory: ‘usr/src/app/capif_onboarding/capif_api_details.json’

I’m calling the prepare.sh script in the Dockerfile like shown in the dummy NetApp:

However, I have the impression that the command is ignored:

Any idea? As suggested there, I saved the script with Unix end of line style.

[sxarism]

Hi Charles.
First of all, to see that the capif registration was made successfully, you can access the mongodb express portal in localhost:8082 (or the IP of the machine hosting the CAPIF).
There, if the registration is successful, there will be a database named “capif”, then a table named “users” and inside you can see the info of your netapp.

Concerning the second one, have you create an empty folder “capif_onboarding” in your netapp?
If you see, in the dummy netapp, inside the python container, the certain empty folder exists.

[Charles_Bailly]

Hi Stavros,

Ok, I will check.

Yes, I had created the capif_onboarding folder before, as shown in the file structure screenshot of my initial post. It is located at the same level of the Dockerfile, capif_registration and prepare.sh script.
The only difference with the initial screenshots is that I’m using /usr/src/app instead of ./ as path for the folder certificates in order to mimic the dummy netapp.

[Charles_Bailly]

I started from scratch on my workstation and got the same result so far. Here is what the database looks like:

When clicking on the user collection:

Overall, no users collection and the infos about the netapp are not there. It seems that the registration/onboarding was not performed (no file created in the capif_onboarding folder, but I have the certificates on app/core/certificates/ on the NEF side).

Edit: I replaced the CMD command in the Dockerfile by a RUN command and the script is now called. However, I have a NewConnectionError as it could not find the capifcore host :

What I tried:

1. Complete the /etc/hosts as shown in the dummy netapp. Got an error telling that the file system is read-only.
2. Add an extra host directly in the docker-compose. Does not seems to do the trick.

I will investigate the n°1 as it looks like a permission issue (even if I already lauch commands in administrator mode).