Skip to content

Latest commit

 

History

History
137 lines (93 loc) · 6.26 KB

howto-install-iot-edge.md

File metadata and controls

137 lines (93 loc) · 6.26 KB

How to Setup a Industrial IoT Edge Gateway

Home

Pre-configured Industrial IoT Edge Gateways

The industrial assets (machines and systems) are connected to Azure through modules running on an Azure IoT Edge industrial gateway.

You can purchase industrial gateways compatible with IoT Edge. Please see our Azure Device Catalog for a selection of industrial-grade gateways. Alternatively, you can setup a local VM.

Manual Industrial IoT Edge Gateway Installation

Create an IoT Edge Instance and Install the IoT Edge Runtime

You can also manually create an IoT Edge instance for an IoT Hub and install the IoT Edge runtime following the IoT Edge setup documentation. The IoT Edge Runtime can be installed on Linux or Windows.

Install the Industrial Modules

The Azure Industrial IoT deployment script will setup IoT Edge Layered Deployments for each Industrial Module. These Layered Deployments will be automatically applied to any IoT Edge instance that contains the following Device Twin JSON tags.

  1. Go to the Azure Portal page and select your IoT Hub

  2. Open the Device Twin configuration JSON under IoT Edge -> [your IoT Edge device] -> Device Twin

  3. Insert the following tags:

  • For Linux (with Linux Containers), set the "os" property to "Linux":
...
},
"version": 1,
"tags": {
    "__type__": "iiotedge",
    "os": "Linux"
},
"properties":
...
  • For Windows (with Linux or Windows Containers), set the "os" property to "Windows":
...
},
"version": 1,
"tags": {
    "__type__": "iiotedge",
    "os": "Windows"
},
"properties":
...

The tags can also be created as part of an Azure Device Provisioning (DPS) enrollment. An example of the latter can be found in /deploy/scripts/dps-enroll.ps1.

Unmanaged Industrial IoT Edge

Layered deployments will make sure that your edge devices will always contain the modules that work with your platform deployment. This includes keeping module versions aligned with the platform version.

However, sometimes it is desirable to not have layered deployments manage your Gateway fleet. Instead you might want to manage the content of the edge gateway yourself. To prevent the platform from creating layered deployments on your IoT Edge Gateway, but still have the Gateway participate in the Platform, you can define a unmanaged tag in addition to the tags above, like so:

...
"tags": {
    "__type__": "iiotedge",
    // ...
    "unmanaged": true
}
...

This will cause that no modules are automatically deployed and thus you must deploy all modules using a module deployment manifest via Az, or Portal.

Module Versions

By default, the same Docker container image version tag from mcr.microsoft.com is deployed that corresponds to the corresponding micro-service's version.

If you need to point to a different Docker container registry or image version tag, you can configure the source using environment variables PCS_DOCKER_SERVER, PCS_DOCKER_USER, PCS_DOCKER_PASSWORD, PCS_IMAGES_NAMESPACE and PCS_IMAGES_TAG, for example in your .env file (which can also be set during deployment), then restart the edge management or all-in-one service.

Special Cases for Windows Networking Configuration

When running the Industrial IoT Edge modules in host (transparent) network mode, the container must be on the transparent host network and will require IP addresses assignment if no DNS server is avialable on that network.

  • Ensure Hyper-V is enabled in the host OS

  • Create a new virtual switch named host and attach it to a network containing the industrial assets you want to connect to (e.g. "Ethernet 2").

    New-VMSwitch -name host -NetAdapterName "<Adapter Name>" -AllowManagementOS $true

  • To make sure the container is assigned an IP address it can either obtain:

    1. A Dynamic IP address from a local DHCP server accessible from the host's network interface associated to the container's host network

    2. A Static IP address assigned on the container create options statement

      In order to allow static IP address assignment on a Windows container, the docker network requires to be created having the the subnet specified identical to the host's interface

      docker -H npipe:////.//pipe//iotedge_moby_engine network create -d transparent -o com.docker.network.windowsshim.interface="Ethernet 2" -o com.docker.network.windowsshim.networkname=host --subnet=192.168.30.0/24 --gateway=192.168.30.1 host

Troubleshooting

To troubleshoot your IoT Edge installation follow the official IoT Edge troubleshooting guide

When device discovery operations fail also make sure to validate:

Host network

Linux containers:

docker network ls
    NETWORK ID          NAME                DRIVER              SCOPE
    beceb3bd61c4        azure-iot-edge      bridge              local
    97eccb2b9f82        bridge              bridge              local
    758d949c5343        host                host                local
    72fb3597ef74        none                null                local

Windows containers:

docker -H npipe:////.//pipe//iotedge_moby_engine network ls
    NETWORK ID          NAME                DRIVER              SCOPE
    8e0ea888dbd4        host                transparent         local
    f3390c998f90        nat                 nat                 local
    6750449db22d        none                null                local

Other Options