Docker images for i open source manangement.
Currently, we have following 4 dockers in this repo.
-
ibmirpm
- This docker could help you to setup RPM env on target IBM i system.
-
ibmibld39
- This docker help you to setup build essential packages, especially for python 3.9 on target IBM i system.
-
ibmiml39
-
This docker helps you to setup following Machine Learning frameworks(based on python3.9) on target IBM i system.
- pybind11 cython six threadpoolctl joblib pyzmq ninja seaborn
- Pillow
- numpy
- scipy
- scikit-learn
- jupyter jupyterlab
- matplotlib
- pandas
-
NOTE: The packages after numpy would take long time depends on your systems. If failed or too long time without any reponse. Don't worry, try to install them directly on IBM i using
python3.9 -m pip install <packageName>
.
-
-
mlserver
- This docker help you to setup Machine Learning Server(mlserver==1.1.0) on IBM i system.
- Following command could help you build out the docker images.
docker build -t <your_repo>/<your_image_name> .
- We also have pre-build images available for your use dicrectly from docker.io. You can just download and use them by docker pull as below:
- docker.io/zheddie/ibmirpm
- docker.io/zheddie/ibmibld39
- docker.io/zheddie/ibmiml39
- docker.io/zheddie/ibmimlserver
-
Following LPP are required tobe installed:
- 5770SS1 Option 33 Portable App Solutions Environment(PASE)
- 5770SS1 Option 13 System Openness Includes
- 5770SC1 Option 1 OpenSSH, OpenSSL, zlib
-
Following PTFs are required
- i 7.3: SI80883
- i 7.4: SI80885
- i 7.5: SI80884
-
To make the OpenSSH work, we need to start the *SSHD server before we access it. Following command could be used to start the *SSHD in 5250 session.
STRTCPSVR SERVER(*SSHD)
-
Non-password access is recommanded, or else, you would be prompt for any remote acess.
ssh-copy-id <your_account>@<target_ibmi_system>
-
To make you easily access all utilities in RPM, PATH env could be set as below when ever you access IBM i through OpenSSH.
export PATH=/QOpenSys/pkgs/bin:~/gcc10bin:$PATH
export OBJECT_MODE=64
doccker run -it --rm -v <your_home_dir>/.ssh/id_rsa:/root/.ssh/id_isa -e ICONNECT=<your_account>@<target_ibmi_system> <your_repo>/<your_image_name>