awsdocs / aws-greengrass-developer-guide

Open source version of the AWS IoT Greengrass V1 Developer Guide

NOTICE

This repository is archived, read-only, and no longer updated. For more information, read the announcement on the AWS News Blog.

You can find up-to-date AWS technical documentation on the AWS Documentation website, where you can also submit feedback and suggestions for improvement.

aws-greengrass-developer-guide's Issues

IDT MQTT Test Fails because of TimeOut Errors

Hi,

When running the IDT, for the MQTT test group, it fails because of a timeout error. What can I do to solve it?

Where can I download the lambda function

It seems the lambda function greengrassObjectClassification is missing from this tutorial

Got a not supported error on a pi-zero but was able to install manually

Code 1, Message: Platform Linux-4.19.75+-armv6l-with-debian-10.1 not supported

Generic Greengrass IoT policy using policy variables

The Minimal Greengrass IoT policy documented in https://github.com/awsdocs/aws-greengrass-developer-guide/blame/master/doc_source/device-auth.md#L112 provides a sample policy that requires you to hard-code the Thing Names and Greengrass Group IDs into the policy statements.

It would be very helpful to add an example of a generic policy that uses IoT policy variables like ${iot:Connection.Thing.ThingName} and ${iot:Connection.Thing.Attributes[greengrassGroupId]} to avoid hard-coding the Thing Names and Greengrass Group IDs. Such a policy could be shared by all deployed Greengrass Cores, eliminating the need to deploy hundreds or thousands of separate policies.

I have been developing such a policy myself recently and it's very difficult to get it right. As a developer there is no visibility to what the current values of policy variables are and debugging is a lot of trial-and-error work. Therefore it would be quite valuable to have an AWS-provided example as a reference.

Docs for IoT Lambda/Greengrass out of date

https://docs.aws.amazon.com/greengrass/latest/developerguide/config-lambda.html

"In the AWS IoT console, under Greengrass, choose Groups, and then choose the group that you created in Module 2."

There is no Groups - the entries are:

Getting started
Core devices
Components
Deployments

Under Manage there is Thing Groups, but there is no way to add a Lamba. Even under Actions.

The docs seem very much out of date with the actual web UI.

The greengrass:Discover permission is not included in the example policy

We spent a lot of time debugging why our IoT policy, based on the "minimal AWS IoT Policy for Greengrass" (https://github.com/awsdocs/aws-greengrass-developer-guide/blob/master/doc_source/device-auth.md) did not allow our Docker services to discover the Greengrass endpoint.

The reason turned out to be that greengrass:Discover is not included in the minimal example policy. It includes greengrass:GetConnectivityInfo, so we thought that would be enough to allow the services to perform discovery. We completely missed that greengrass:Discover was even mentioned in the documentation.

Since many developers probably copy and adapt the example policy, I would suggest either adding the greengrass:Discover permission to it, or at least adding a note reminding that you will also need greengrass:Discover if you want to perform discovery.

Note that we are using a shared IoT policy that covers both our Greengrass Core and the Docker services running as "devices" in the Greengrass Group. I don't know if this is a normal use case, but we certainly don't want to create even more separate IoT policies for all the Docker containers running on every device.

PS. To add some difficulty, greengrass:Discover is also not included in this documentation: https://docs.aws.amazon.com/IAM/latest/UserGuide/list_awsiotgreengrass.html#awsiotgreengrass-actions-as-permissions. So it's pretty difficult to spot it when you don't even realize it exists.

Access denied trying to update Greengrass Core with CLI

I am trying the update the Greengrass Core on a device. Upon running the command:

aws greengrass create-software-update-job     --update-targets-architecture armv7l     --update-targets arn:aws:iot:eu-west-1:MYACCOUNTID:thing/raspberrydnb1_Core     --update-targets-operating-system raspbian     --software-to-update core     --s3-url-signer-role arn:aws:iam::MYACCOUNTID:role/service-role/Greengrass_ServiceRole     --update-agent-log-level WARN

I get an error:

An error occurred (InvalidRequestException) when calling the
CreateSoftwareUpdateJob operation: Permission denied when attempting
to assume role arn:aws:iam::MYACCOUNTID:role/GreenGrassUpdateRole.

My trust policy on the signer role is like:

{
      "Effect": "Allow",
      "Principal": {
        "AWS": "arn:aws:iam::MYACCOUNTID:root",
        "Service": "greengrass.amazonaws.com"
      },
      "Action": "sts:AssumeRole"
    }

By the way, this command works:

aws sts assume-role --role-arn arn:aws:iam::MYACCOUNTID:role/GreenGrassUpdateRole --role-session-name example-role-assume

greengrass is not authorized to assume the service role associated with this account

Hi Guys,
when trying to follow your instructions and after trying to deploying the group rule i get the above message.. any clue how to overcome it?
kind regards,
Christos

ML Inference on Raspberry Pi requires 2.5 A power source

The inference instructions should note that running deep learning frameworks on a Raspberry Pi requires a 2.5A power source. If you use a lower rated power supply, the Pi will likely reboot whenever you try to perform inference by running the forward method on the network, at least for deep networks that you see with image use cases. I saw this behavior with MxNet, the DLR, and TensorFlow on a Pi 3B+.

aws iot greengrass on jetson nano - local resource access failed

Hi,

I'm trying to implement aws greengrass on jetson nano I'm following below aws documentation;
https://docs.aws.amazon.com/greengrass/latest/developerguide/lra-console.html

But i can't able to proceed or having issue read/write a file from cloud to edge. Im having spending more a days, didn't find and can't able fix this issue. Please help me out!!!
Here is my log while lambda running;

[2019-07-01T15:57:47.009+05:30][INFO]-lambda_runtime.py:331,Caught signal 15. Stopping runtime.
[2019-07-01T15:57:59.286+05:30][INFO]-lambda_runtime.py:114,Running [arn:aws:lambda:us-east-1:xxxxxxxxxx:function:TestLRA:7]
[2019-07-01T15:57:59.286+05:30][INFO]-ipc_client.py:170,Getting work for function [arn:aws:lambda:us-east-1:xxxxxxxxxxx:function:TestLRA:7] from http://localhost:8000/2016-11-01/functions/arn:aws:lambda:us-east-1:xxxxxxxxxxx:function:TestLRA:7/work

Could not find suitable distribution for Requirement.parse('dlr==1.0')

Hi there,
Could you help me to take a look? Thank you!

Documentation:
Configure Optimized Machine Learning Inference Using the AWS Management Console

Problem Description:

1. Operating system: Raspbian for Raspberry Pi 3B+

2. Problem at Step 2.7 Install the Neo deep learning runtime

3. The error I got: 'No local packages or working download links found for dlr==1.0
   error: Could not find suitable distribution for Requirement.parse('dlr==1.0')'

4. I also tried installation using the below two ways but neither worked.

• neo-ai-dlr repo
  https://github.com/neo-ai/neo-ai-dlr/tree/master/install/dlr-1.0-py2.py3-armv7l
• pip install dlr
  https://pypi.org/project/dlr/