Configuration documentation

[alfonsoegio]

Please, document all the configuration management / files to configure ports / IPs (specially the slice-manager API url); integration documentation with dashboard also.

[ubiteixeira]

@alfonsoegio I've added new information on the README and also added a dedicated README for the Gravitee located on gravitee/README.md

[alfonsoegio]

Sorry, can you point me to the part of the documentation explaining how the Slice Manager base URL is configured? performing:
$ rgrep slave1
gravitee/data/Slice Manager API/api.json "target":"http://5gcityslave1.i2cat.net:8989/api/v0.1/",

[ubiteixeira]

Since the import files deal with many configurations and Gravitee provides an intuitive UI, what we did was to provide documentation to help manage the APIs through Gravitee's UI, avoiding dealing directly with the JSONs.

https://github.com/5GCity/5GCity-AAA/blob/master/gravitee/README.md

However if you feel it's more productive to edit the JSON directly I can manage to describe some of its fields.

If the administrator needs to change some of the API configurations it can easily export the API overwriting the old configuration.

[alfonsoegio]

machine still raises a "Management API unreachable or error occurs, please check logs" on port 80, it seems to try to access localhost:8083 instead of the url I configured in /etc/config.ini:

[gravitee]
username='admin'
password='admin'
base_url='http://5gcity-slave1.i2cat.net:8083/management'

[alfonsoegio]

Also when trying to execute: python main.py gravitee --import_api "Slice Manager API"

I'm getting an error; maybe this has to do with performing the setup procedure many times, can you document how to stop / regenerate the containers, in order to get config changes properly applied?

[ubiteixeira]

Please get the new version of the code and try again please. I refactored the configurations and added a function to validate your mmap counts which I believe it was the problem.

https://www.elastic.co/guide/en/elasticsearch/reference/current/vm-max-map-count.html

You can also check on the README the port mapping of each service. Now the Gravitee UI management will be placed on 8092 by default.

[alfonsoegio]

Should I perform a python main.py docker --stop && docker --shutdown before? If that's the case, please document it also

[ubiteixeira]

Yes please use the stop and shutdown commands before starting and importing both the API and Keycloak realm

[alfonsoegio]

After stopping and shutting down, executing "python main.py docker --start": and waiting for message "Admin console listening on http://127.0.0.1:9990" when running "python main.py keycloak --import_realm 5gcity" I'm obtaining the message:

docker: Error response from daemon: Conflict. The container name "/keycloak_exporter" is already in use by container "d8bea3bc5c32c5c512750bfb2e2f45452ffac6c71cbcc9c56a28b39cdcfa59a8". You have to remove (or rename) that container to be able to reuse that name.
See 'docker run --help'.

[ubiteixeira]

This is because the keycloak exporter launches a new container by the name of keycloak_exporter which is still running. Please remove the container and try again the export. We're still working on a way to make this process cleaner.

[alfonsoegio]

OK, now I've reached the point: python main.py gravitee --import_api "Slice Manager API"

runs without error messages but, accessing 8092 port on host machine still fails: "Management API unreachable or error occurs, please check logs" trying to connect to http://localhost:8083 error.

Maybe the error has to do with the fact that MGMT_API_URL is pointing to localhost in the compose-aaa.yml configuration and if that's the case it would be necessary to clarify which URLs on this configuration file correspond to internal management of services and which should be considered external (for example the ones that are going to be used by user's browser).

[alfonsoegio]

If I change in compose-aaa.yml localhost:8083 for 5gcity-slave1.i2cat.net:8083 (to prevent the fact that my browser thinks that the 8083 service runs on my own machine) when performing "python main.py gravitee --import_api "Slice Manager API"" I'm getting an error:

Traceback (most recent call last):
  File "main.py", line 40, in <module>
    get_args()
  File "main.py", line 36, in get_args
    COMMANDS[command].execute(args)
  File "/home/ubuntu/5GCity/5GCity-AAA/commands/gravitee.py", line 21, in execute
    getattr(mcs, arg)(value)
  File "/home/ubuntu/5GCity/5GCity-AAA/commands/gravitee.py", line 31, in import_api
    g.import_api(args[0])
  File "/home/ubuntu/5GCity/5GCity-AAA/gravitee/gravitee.py", line 144, in import_api
    api_id = __api__()
  File "/home/ubuntu/5GCity/5GCity-AAA/gravitee/gravitee.py", line 116, in __api__
    return r.json()['id']
KeyError: 'id'

Seems that gravitee UI deployment does not work apart from being used from the same computer?

[ubiteixeira]

Sorry for this. I included just know an error feedback regarding the gravitee commands and also an URL option so you can point the Management API you want to use, e.g.,

python main.py gravitee --import_api "Slice Manager API" --url "http://5gcity-slave1.i2cat.net:8083/management/"

[alfonsoegio]

Hello again:

$ python main.py gravitee --import_api "Slice Manager API" --url "http://5gcity-slave1.i2cat.net:8083/management/"
Command error with code 400 and message {
  "message" : "The api context path [/slicemanagerapi/] already exists.",
  "http_status" : 400
}