AWS Greengrass V2批量配置IoT策略部署失败问题排查

我在已开放所需端口的EC2实例上配置AWS Greengrass V2批量部署，证书已生成且核心设备完成注册。为满足生产安全要求，采用了Greengrass核心最小IoT策略，具体策略如下：

{
  "Statement": [
    {
      "Action": [
        "iot:Publish",
        "iot:Subscribe",
        "iot:Receive",
        "iot:Connect"
      ],
      "Effect": "Allow",
      "Resource": [
        "arn:aws:iot:eu-central-1:123123123123:topic/data/${iot:Connection.Thing.ThingName}/*",
        "arn:aws:iot:eu-central-1:123123123123:topic/cmd/${iot:Connection.Thing.ThingName}/*"
      ]
    },
    {
      "Action": [
        "iot:Connect"
      ],
      "Effect": "Allow",
      "Resource": "arn:aws:iot:eu-central-1:123123123123:client/${iot:Connection.Thing.ThingName}*"
    },
    {
      "Action": [
        "iot:Subscribe"
      ],
      "Effect": "Allow",
      "Resource": [
        "arn:aws:iot:eu-central-1:123123123123:topicfilter/$aws/things/${iot:Connection.Thing.ThingName}*/jobs/*",
        "arn:aws:iot:eu-central-1:123123123123:topicfilter/$aws/things/${iot:Connection.Thing.ThingName}*/shadow/*"
      ]
    },
    {
      "Action": [
        "iot:Receive",
        "iot:Publish"
      ],
      "Effect": "Allow",
      "Resource": [
        "arn:aws:iot:eu-central-1:123123123123:topic/$aws/things/${iot:Connection.Thing.ThingName}*/greengrass/health/json",
        "arn:aws:iot:eu-central-1:123123123123:topic/$aws/things/${iot:Connection.Thing.ThingName}*/greengrassv2/health/json",
        "arn:aws:iot:eu-central-1:123123123123:topic/$aws/things/${iot:Connection.Thing.ThingName}*/jobs/*",
        "arn:aws:iot:eu-central-1:123123123123:topic/$aws/things/${iot:Connection.Thing.ThingName}*/shadow/*"
      ]
    },
    {
      "Action": [
        "greengrass:ResolveComponentCandidates",
        "greengrass:Get*",
        "greengrass:List*",
        "greengrass:Describe*",
        "greengrass:Resolve*",
        "greengrass:PutCertificateAuthorities"
      ],
      "Effect": "Allow",
      "Resource": "*"
    },
    {
      "Action": "iot:AssumeRoleWithCertificate",
      "Effect": "Allow",
      "Resource": "arn:aws:iot:eu-central-1:123123123123:rolealias/TerraformGreengrassCoreTokenExchangeRoleAlias"
    }
  ],
  "Version": "2012-10-17"
}

核心设备无法接收Thing Group的部署，且频繁断开连接，日志片段如下：

[...]
2023-01-05T08:58:18.602Z [DEBUG] (pool-2-thread-37) com.aws.greengrass.mqttclient.AwsIotMqttClient: Subscribing to topic. {clientId=TestCustomerCoreDevice, qos=AT_LEAST_ONCE, topic=$aws/things/TestCustomerCoreDevice/jobs/12312397-1d2d-1d2d-1d2d-01de629ddcf2/namespace-aws-gg-deployment/update/rejected}
com.aws.greengrass.mqtt.bridge.clients.MQTTClient: Unable to connect. Will be retried after 120 seconds
[...]

将订阅权限放宽到"arn:aws:iot:eu-central-1:123123123123:*"可正常工作，但不符合生产安全要求。推测问题与$aws前缀的topicfilter资源有关，但无法定位具体原因。权限放宽后可正常订阅data/TestCustomerCoreDevice/test主题，求最小权限下的解决办法。

问题出在IoT策略中topicfilter的资源匹配规则，原策略里的${iot:Connection.Thing.ThingName}*多了一个通配符*，导致无法精确匹配设备的Thing名称（比如你的设备是TestCustomerCoreDevice，没有后缀）。

需要调整策略中以下两处的资源：

1. **订阅权限（iot:Subscribe）**的topicfilter资源，去掉ThingName后的*
2. **收发权限（iot:Receive/iot:Publish）**的topic资源，同样去掉ThingName后的*

调整后的完整策略如下：

{
  "Statement": [
    {
      "Action": [
        "iot:Publish",
        "iot:Subscribe",
        "iot:Receive",
        "iot:Connect"
      ],
      "Effect": "Allow",
      "Resource": [
        "arn:aws:iot:eu-central-1:123123123123:topic/data/${iot:Connection.Thing.ThingName}/*",
        "arn:aws:iot:eu-central-1:123123123123:topic/cmd/${iot:Connection.Thing.ThingName}/*"
      ]
    },
    {
      "Action": [
        "iot:Connect"
      ],
      "Effect": "Allow",
      "Resource": "arn:aws:iot:eu-central-1:123123123123:client/${iot:Connection.Thing.ThingName}*"
    },
    {
      "Action": [
        "iot:Subscribe"
      ],
      "Effect": "Allow",
      "Resource": [
        "arn:aws:iot:eu-central-1:123123123123:topicfilter/$aws/things/${iot:Connection.Thing.ThingName}/jobs/*",
        "arn:aws:iot:eu-central-1:123123123123:topicfilter/$aws/things/${iot:Connection.Thing.ThingName}/shadow/*"
      ]
    },
    {
      "Action": [
        "iot:Receive",
        "iot:Publish"
      ],
      "Effect": "Allow",
      "Resource": [
        "arn:aws:iot:eu-central-1:123123123123:topic/$aws/things/${iot:Connection.Thing.ThingName}/greengrass/health/json",
        "arn:aws:iot:eu-central-1:123123123123:topic/$aws/things/${iot:Connection.Thing.ThingName}/greengrassv2/health/json",
        "arn:aws:iot:eu-central-1:123123123123:topic/$aws/things/${iot:Connection.Thing.ThingName}/jobs/*",
        "arn:aws:iot:eu-central-1:123123123123:topic/$aws/things/${iot:Connection.Thing.ThingName}/shadow/*"
      ]
    },
    {
      "Action": [
        "greengrass:ResolveComponentCandidates",
        "greengrass:Get*",
        "greengrass:List*",
        "greengrass:Describe*",
        "greengrass:Resolve*",
        "greengrass:PutCertificateAuthorities"
      ],
      "Effect": "Allow",
      "Resource": "*"
    },
    {
      "Action": "iot:AssumeRoleWithCertificate",
      "Effect": "Allow",
      "Resource": "arn:aws:iot:eu-central-1:123123123123:rolealias/TerraformGreengrassCoreTokenExchangeRoleAlias"
    }
  ],
  "Version": "2012-10-17"
}

补充说明 #

• Greengrass核心设备订阅部署相关的Job主题时，使用的是精确的Thing名称（无后缀），原策略中的${iot:Connection.Thing.ThingName}*会匹配带后缀的名称，导致权限不匹配
• 保留client资源中的*是合理的，因为Greengrass可能会在clientId后添加随机后缀用于重连
• 调整后策略仍保持最小权限，仅允许设备访问自身相关的IoT主题和Greengrass资源

内容的提问来源于stack exchange，提问作者DK_kbc