Make sure you have the registry credentials handy. If you have not received the registry credentials, you need to contact your Sales representative and request a set of credentials be generated for you. Then type the following in a Powershell or Command Line terminal (on Windows) or shell terminal (on Linux):
docker login aiflprodweacr.azurecr.io -u <username> -p <password>
docker pull aiflprodweacr.azurecr.io/datamanager:latest
For more details about how to use Data Manager see this documentation page.
Running on the same machine as AI Fabric v2020.7 or later
Data Manager standalone container cannot run on the same machine as AI Fabric v2020.7 or later. It can run on the same machine as AI Fabric v2020.4, or on any other Windows 10 or Linux machine.
To launch Data Manager, use the following command:
docker run -d -p <port_number>:80 -v "<path_to_working_folder>:/app/data" aiflprodweacr.azurecr.io/datamanager:latest --license-agreement accept
Open a web browser and enter the following URL:
If you are using the browser on a different machine, replace
localhost with the IP address of the machine where the datamanager container is running.
If an ML Skill deployed in AI Fabric on premises does not use a valid HTTPS certificate, you can use the CLI option to whitelist the root of your self-signed certificate.
--custom-root-cert CLI argument should contain the path towards a custom certificate.
The certificate needs to be in a PEM format. As long as this requirement is met, the file extension is insignificant.
The certificate has to exist inside the docker container, so it has to be mounted. Hence, mount the cert file inside the container and then specify the path to it:
docker run -d -p <port_number>:80 -v "<path_to_working_folder>:/app/data" -v "<path_to_certificate_file>":/custom.cer aiflprodweacr.azurecr.io/datamanager:latest --license-agreement accept --custom-root-cert="/custom.cer"
The root of the self-signed certificate, in this case
custom.cer, has to be the same in the first argument as well as in the second. If one is changed, the other needs to be changed too.
If you need to set up Data Manager on a machine with no internet access (airgapped), you need to run the above commands on some other machine which does have internet access. Then you need to save the container as a .tar file, copy the file over to the airgapped machine, and then load it. This is done using the docker save and docker load commands described in Docker documentation.
So on the machine connected to the internet, you need to first install Docker, then, after running the docker login and docker pull commands above, you need to run the command
docker save -o datamanager-latest.tar aiflprodweacr.azurecr.io/datamanager:latest
Then you need to copy the .tar file to the airgapped machine, and then run this command in the same folder where the .tar file was saved:
docker load --input datamanager-latest.tar
Be aware that the tar file will be large, it will have a few gigabytes.
Updated 2 days ago