influxdb-export application service resetting the values and not taking new values from consul registry

[mohammedimrankasab]

Hi

we are currently working on containerizing the influxdb-export app service and run it using the docker-compose.

The service is running as expected but there is an observation that whenever we update any custom values like the mqtt broker address using the consul registry it is not reflected and if we restart the application service container then the old value from the configuration.yaml is getting overridden hence not accepting any new values through the consul UI.

  app-influxdb-export:
    command:
    - /app-influxdb-export
    - -cp=consul.http://edgex-core-consul:8500
    - --registry
    - -skipVersionCheck=true
    container_name: edgex-app-influxdb-export
    depends_on:
      consul:
        condition: service_started
      core-data:
        condition: service_started
      security-bootstrapper:
        condition: service_started
    entrypoint:
    - /edgex-init/ready_to_run_wait_install.sh
    environment:
      EDGEX_SECURITY_SECRET_STORE: "true"
      PROXY_SETUP_HOST: edgex-security-proxy-setup
      SECRETSTORE_HOST: edgex-vault
      SERVICE_HOST: edgex-app-influxdb-export
      STAGEGATE_BOOTSTRAPPER_HOST: edgex-security-bootstrapper
      STAGEGATE_BOOTSTRAPPER_STARTPORT: "54321"
      STAGEGATE_DATABASE_HOST: edgex-redis
      STAGEGATE_DATABASE_PORT: "6379"
      STAGEGATE_DATABASE_READYPORT: "6379"
      STAGEGATE_PROXYSETUP_READYPORT: "54325"
      STAGEGATE_READY_TORUNPORT: "54329"
      STAGEGATE_REGISTRY_HOST: edgex-core-consul
      STAGEGATE_REGISTRY_PORT: "8500"
      STAGEGATE_REGISTRY_READYPORT: "54324"
      STAGEGATE_SECRETSTORESETUP_HOST: edgex-security-secretstore-setup
      STAGEGATE_SECRETSTORESETUP_TOKENS_READYPORT: "54322"
      STAGEGATE_WAITFOR_TIMEOUT: 60s
    hostname: edgex-app-influxdb-export
    image: edgexfoundry/app-influxdb-export:10.0.0-dev
    networks:
      edge-network: null
    ports:
    - mode: ingress
      host_ip: 127.0.0.1
      target: 59780
      published: "59780"
      protocol: tcp
    read_only: true
    restart: always
    security_opt:
    - no-new-privileges:true
    user: 2002:2001
    volumes:
    - type: volume
      source: edgex-init
      target: /edgex-init
      read_only: true
      volume: {}
    - type: bind
      source: /etc/localtime
      target: /etc/localtime
      read_only: true
      bind:
        create_host_path: true
    - type: bind
      source: /tmp/edgex/secrets/app-influxdb-export
      target: /tmp/edgex/secrets/app-influxdb-export
      read_only: true
      bind:
        selinux: z
        create_host_path: true

Note: No changes in the code or the configuration.yaml, The only thing is using this code a docker image is built and used as a docker container.

Need help on why the configuration is getting overwritten with default values from configuration.yaml and not taking the latest value from the consul UI.

Thanks

[cloudxxx8]

you may check the logs to get more information
which key path did you try to modify?
can the Service.StartupMsg be changed?

[mohammedimrankasab]

Thanks for the reply @cloudxxx8

Any keys under the path Key/Values/edgex/v3/app-influxdb-export/MqttConfig is not getting updated.

I tried to change the AuthMode: "usernamepassword" to "none" and change the BrokerAddress

yes, I can update the StartupMsg and also can see the new message after restarting the container

level=INFO ts=2024-03-08T06:48:36.014365822Z app=app-influxdb-export source=service.go:244 msg="App service export to InfluxDB has Started **again**"

There are no error logs which I can see on the app service container but on the consul container I can see the below logs

2024-03-08T06:48:02.282Z [ERROR] agent.http: Request error: method=POST url=/v1/internal/acl/authorize?dc=dc1 from=192.168.160.1:35464 error="ACL not found"
2024-03-08T06:48:02.498Z [ERROR] agent.http: Request error: method=POST url=/v1/internal/acl/authorize?dc=dc1 from=192.168.160.1:35464 error="ACL not found"
2024-03-08T06:48:02.662Z [ERROR] agent.http: Request error: method=GET url=/v1/kv/edgex/v3/app-influxdb-export/MqttConfig/BrokerAddress?dc=dc1 from=192.168.160.1:35474 error="ACL not found"

Below entries are already updated on the docker-compose file
EDGEX_ADD_REGISTRY_ACL_ROLES: "app-influxdb-export"
EDGEX_ADD_KNOWN_SECRETS: redisdb[app-influxdb-export],message-bus[app-influxdb-export]
EDGEX_ADD_SECRETSTORE_TOKENS: "app-influxdb-export"

the only problem is with the MqttConfig section

[cloudxxx8]

it looks like only custom config cannot be updated from Config Provider
please open a bug in app-functions-sdk repository

[mohammedimrankasab]

sure. Thanks for the clarification. Here is the issue raised

[lenny-goodell]

Note that this example doesn't watch the custom section MqttConfig for changes, so a restart is always required after changes in Consul to this custom section.

I did confirm that the custom section MqttConfig is not being loaded from Consul on restart, so this is bug has been recreated, but it is in the example, not the SDK.

It seems to be checking for the wrong name in Consul.

msg="Checking if custom configuration ('MqttSecretConfig') exists in Configuration Provider"
msg="Loading configuration file from res/configuration.yaml"
msg="Loaded custom configuration from File (0 envVars overrides applied)"
msg="Custom Config loaded from file and pushed to Configuration Provider "

Simply change the section name used in this line to MqttConfig
https://github.com/edgexfoundry/edgex-examples/blob/main/application-services/custom/influxdb-export/main.go#L37

Once you do this you will see the following in the logs
msg="Checking if custom configuration ('MqttConfig') exists in Configuration Provider"
msg="Loaded custom configuration from Configuration Provider, no overrides applied"

I am moving the bug issue to the edgex-examples repo