Greengrass Lambdamanager fails when I set up a LegacySubscriptionRouter

Question

Hi, I'm setting up GreengrassV2 and the Modbus adapter on a RPI3b.
I'm running

* aws.greengrass.Modbus: 2.1.8
* aws.greengrass.LegacySubscriptionRouter: 2.1.11
* aws.greengrass.LambdaManager: 2.3.2

I'm using more or less default for everything, I just added a USB modbus adapter as a device by merging the following through the UI

```
{
  "lambdaExecutionParameters": {
    "EnvironmentVariables": {
      "ModbusLocalPort": "/dev/ttyUSB0"
    }
  },
  "containerParams": {
    "memorySize": 525312,
    "mountROSysfs": false,
    "volumes": {},
    "devices": {
      "0": {
        "path": "/dev/ttyUSB0",
        "permission": "rw",
        "addGroupOwner": true
      }
    }
  }
}
```
and set up the LegacySubscriptionRouter to read from the topic
```
{
    "subscriptions": {
      "aws-greengrass-modbus": {
        "id": "aws-greengrass-modbus",
        "source": "component:aws.greengrass.Modbus",
        "subject": "modbus/adapter/response",
        "target": "cloud"
      }
    }
  }
```

Howver this breaks the LambdaManager:
```
2023-11-15T10:24:21.627Z [WARN] (Serialized listener processor) com.aws.greengrass.lambdamanager.LambdaManager: service-invalid-state-error. Invalid reported state. {serviceName=aws.greengrass.LambdaManager, currentState=BROKEN, newState=ERRORED}
2023-11-15T10:24:21.635Z [ERROR] (Serialized listener processor) com.aws.greengrass.lambdamanager.LambdaManager: service-errored. {serviceName=aws.greengrass.LambdaManager, currentState=BROKEN}
com.aws.greengrass.lambdamanager.system.v1subscription.InvalidArnException: The provided arn is not in proper format, arn: modbus/adapter/response
	at com.aws.greengrass.lambdamanager.system.v1subscription.RouteTable.parseArnType(RouteTable.java:163)
	at com.aws.greengrass.lambdamanager.system.v1subscription.RouteTable.convertConfigToRoute(RouteTable.java:110)
	at com.aws.greengrass.lambdamanager.system.v1subscription.RouteTable.addRoute(RouteTable.java:98)
	at com.aws.greengrass.lambdamanager.system.v1subscription.RouteTable.loadRoutesFromConfig(RouteTable.java:59)
	at com.aws.greengrass.lambdamanager.system.RouterLambda.loadSubscription(RouterLambda.java:60)
	at com.aws.greengrass.lambdamanager.LambdaManager.reloadV1Subscription(LambdaManager.java:351)
	at com.aws.greengrass.lambdamanager.LambdaManager.lambda$install$1(LambdaManager.java:147)
	at com.aws.greengrass.config.Topics.childChanged(Topics.java:439)
	at com.aws.greengrass.config.Topics.childChanged(Topics.java:461)
	at com.aws.greengrass.config.Topics.childChanged(Topics.java:461)
	at com.aws.greengrass.config.Topics.childChanged(Topics.java:461)
	at com.aws.greengrass.config.Topics.childChanged(Topics.java:461)
	at com.aws.greengrass.config.Topics.lambda$remove$9(Topics.java:418)
	at com.aws.greengrass.dependency.Context$1.run(Context.java:68)
```

Any tips why? Seems it expects the source to an ARN, but I'm using the LegacySubscriptionRouter from the dev guide: https://docs.aws.amazon.com/greengrass/v2/developerguide/modbus-rtu-protocol-adapter-component.html#modbus-rtu-protocol-adapter-component-output-data

Does it actually need an ARN instead here?

MichaelDombrowski-AWS · Accepted Answer

Hello,

The error shows that you actually configured the "target" with the value "modbus/adapter/response" instead of setting that as the subject and target as cloud.

Please make sure you reset all legacy subscription router configuration by using `"RESET": [""]` in the deployment configuration update and try merging the correct configuration once again. https://docs.aws.amazon.com/greengrass/v2/developerguide/update-component-configurations.html#reset-configuration-update

You can view the current configuration on the device using the local debug console or local Greengrass CLI components. If you inspect the configuration you would see the problem that I'm talking about.

Cheers,

Michael

Haakon · Answer

Seems there's little functionality for monitoring what's happening during a deployment, are there logs for that somewhere? The job says "In progress - Rollout completed", but there's little other info available.

Haakon · Answer

Spot on, thanks! I tried that configuration first, but I had actually set up the correct setting after that, but missed the `""` in the `RESET[""]`.

I'm retrying the deployment right now, but it gets stuck on "Queued", perhaps because there's a broken component? Seen that happen earlier when there's an issue on an already deployed component, any tips on how to get past that?

Greengrass Lambdamanager fails when I set up a LegacySubscriptionRouter

Relevant content