Skip to content

Latest commit

 

History

History
58 lines (37 loc) · 3.64 KB

troubleshooting.md

File metadata and controls

58 lines (37 loc) · 3.64 KB

Azure IoT Edge Hands On Labs

Created and maintained by the Microsoft Azure IoT Global Black Belts

Troubleshooting

This troubleshooting guide contains some tips for troubleshooting your labs.

General tips

Read the instructions VERY carefully. A good number of the 'issues' found in early internal testing existed between the seat and the keyboard :-). The instructions can get very detailed in places, so if something doesn't work, make sure to re-read the instructions.

Make sure you are always working with the latest set of instructions (i.e. refresh your browser). We find and fix stuff regularly.

Docker tips

A few very useful docker commands:

  • docker ps: gives a list of running docker containers
  • docker images : gives a list of locally caches images
  • docker image rm -f <image name> : removes (forceably) a locally cached image (forcing a fresh re-download) where <image name> is the image name in the format of <repository>/<image>:<tag>. You can alternately just specify the image ID (found from 'docker images') instead of the name. Generally you want to make sure the Edge runtime is stopped when you do this (iotedgectl stop)
  • docker container ls -a : gives a list of docker containers (whether running or not)
  • docker container rm -f <container name> : removes (forceably) a docker container and makes the edge runtime recreate it when it next starts. Generally you want to make sure the Edge runtime is stopped when you do this (iotedgectl stop)
  • docker logs -f <container name> : shows (and follows) the logs for a running container. CTRL-C to exit it.

Of all of these, 'docker logs -f <container name>' is your best friend. It shows anything written to the console (i.e. printf, Console.Writeline, etc) from within the container. All of the MSFT provided modules, and the ones you develop in the lab, provide useful information in the logs

Lab troubleshooting tips

General tips

  • check the case on everything! module names, routes, etc are all CASE SENSITIVE. If something is not connecting or data is not flowing, check the case between the module names and the routes.
  • when it comes to troubleshooting, the first stop should be the docker logs for the modules. We can't say this enough :-)

when my IoT Device (python script) tries to connect to Edge, I get a TLS error or some other error

  • Make sure that your GatewayHostName parameter is spelled correctly ('mygateway.local')
  • make sure you can 'ping' the hostname (should resolve to 127.0.0.1)
  • make sure that the edgeHub is actually running and listening on port 8883. run 'docker ps' to confirm (as in the image)

edgeHubrunning

  • finally, look in the edgeHub logs (docker logs -f edgeHub) for any errors

My module(s) are up and running fine (via docker ps and docker logs), but no messages are 'flowing' through them. In other words, messages are going into edgeHub from my python script, but don't seem to get routed anywhere (including IoT Hub)

We've seen a few times where edgeHub seems to just kind-of get 'stuck'. No errors in the log, and you can send data, C2D messages, DM calls, etc through it, but no messages get routed to either modules or IoT Hub. The only fix we've found so far is to delete, and let Edge re-create, the edgeHub container.

To do so:

  • stop your python script/IoT device if it is running
  • stop IoT Edge (sudo systemctl stop iotedge)
  • delete the edgeHub container
    • docker container rm -f edgeHub
  • delete the edgeHub image (probably not necessary, but why not)
    • docker image rm -f mcr.microsoft.com/azureiotedge-hub:1.0
  • restart IoT Edge (sudo systemctl start iotedge)
  • restart your IoT device