UDMI / Docs / UDMIS / Local Docker
This document describes how to setup and run all the UDMI components locally using docker, so no specific install required. All the commands here should be direct cut-and-paste into a suitable shell window.
Note: This is still in ‘alpha’ form. It works, but there’s likely going to be some kinks to work out (likely to do with authentication, because authentication is annoying).
The system requires a valid site_model to run. This overview uses the default udmi_site_model, but
any valid site model will work. To setup the model to use, create a link from site_model in the working
directory to the model of choice. Also need to create
git clone https://github.com/faucetsdn/udmi_site_model.git
ln -s udmi_site_model/ site_model
docker inspect -f ok udminet || docker network create udminet --subnet 192.168.99.0/24
Starts the background UDMIS locally. The output of this command is uninteresting (assuming it works), so see the next section for some sample diagnostic commands. It will take some time (less than 30s) to get up and going before the other commands will work.
docker run -d --rm --net udminet --name udmis -p 8883:8883 \
-v $(realpath site_model):/root/site_model \
-v $PWD/var/tmp:/tmp \
-v $PWD/var/etcd:/root/udmi/var/etcd \
-v $PWD/var/mosquitto:/etc/mosquitto \
ghcr.io/faucetsdn/udmi:udmis-latest udmi/bin/start_local block site_model/cloud_iot_config.json
Some sample commands to inspect and diagnose what’s going on with the UDMIS startup. See the sample UDMIS output for what this might look like. Real diagnosing will require a bit more investigation than shown here, but if you see the same output then you can assume that the system is working!
docker logs udmis 2>&1 | fgrep udmis
fgrep pod_ready var/tmp/udmis.log
ls -l var/tmp/pod_ready.txt
After startup, the site model needs to be registered as per standard UDMI practice. This only needs to be done once per site model (or after any significant changes). The sample registrar output shows what a successful run looks like.
docker run --rm --net udminet --name registrar -v $(realpath site_model):/root/site_model \
ghcr.io/faucetsdn/udmi:validator-latest bin/registrar site_model/cloud_iot_config.json
For initial install testing, it’s recommended to try first with the standard pubber utility. See the sample pubber output for the beginning of what this run looks like.
docker run -d --rm --net udminet --name pubber -v $(realpath site_model):/root/site_model \
ghcr.io/faucetsdn/udmi:pubber-latest bin/pubber site_model/cloud_iot_config.json
Sequencer can be run directly as per normal too. See the sample sequencer output for what the beginning of a successful run looks like.
docker run --rm --net udminet --name sequencer -v $(realpath site_model):/root/site_model \
ghcr.io/faucetsdn/udmi:validator-latest bin/sequencer site_model/cloud_iot_config.json
The resulting output can be extracted from the site model, See the sample report output for what this should look like.
head site_model/out/devices/*/results.md
For development purposes, the following command will build (and push) the requisite docker images. This requires a full UDMI install and appropriate permissions (to push).
for image in udmis validator pubber; do
docker rmi ghcr.io/faucetsdn/udmi:$image-latest; \
bin/container $image push \
&& docker tag $image:latest ghcr.io/faucetsdn/udmi:$image-latest \
&& docker push ghcr.io/faucetsdn/udmi:$image-latest
done