#Containerization

Sometimes customers need a small IRIS instance to do something in the cloud and shut it down, or they need hundreds of containers (i.e. one per end user or one per interface) with small workloads. This exercise came about to see how small an IRIS instance could be. For this exercise we focused on what is the smallest amount of memory we can configure for an IRIS instance. Do you know all the parameters that affect the memory allocated by IRIS ?

We are a group of interface analysts in a healthcare setting, running IRIS for Health 2024, having upgraded over time from an old ENSEMBLE environment.  We have been working on a wish list of development goals, but are having trouble finding the correct, best practice method/path for getting to what we envision.

Current state:  We run IRIS for Health on a Redhat 8 virtual machine.  We run Interoperability productions (IE.cls) out of 6 namespaces.  We have a dev/test box and a production box.

Some time ago, I put together a small example to quickly deploy InterSystems IRIS instances connected via ECP using Docker.

Time passed and, like everything, it needed a bit of a refresh.

Introduction

In InterSystems IRIS 2024.3 and subsequent IRIS versions, the AutoML component is now delivered as a separate Python package that is installed after installation. Unfortunately, some recent versions of Python packages that AutoML relies on have introduced incompatibilities, and can cause failures when training models (TRAIN MODEL statement). If you see an error mentioning "TypeError" and the keyword argument "fit_params" or "sklearn_tags", read on for a quick fix.

Root Cause

• scikit-learn updated to version 1.6.0, deprecating fit_params.

InterSystems IRIS Community Edition HAOS Add-On

Run InterSystems IRIS inside of Home Assistant, as an add-on.  Before you dismiss this article possibly under the guise that this is just a gimmick, Id like you to step back and take a look at how easy it is to launch IRIS based applications using this platform.  If you look at Open Exchange, you will see dozens of dozens of applications worthy of launching while they are basically hung out to dry as gitware, and launchable if you want to get into a laptop battle with containerd or Docker.  With a simple git repo, and a specification, you can now build your app on IRIS, and make it launchable through a marketplace with limited hassle to your end users.  Run it along side Ollama and the LLM/LAM implementations, expose anything in IRIS as a sensor or expose an endpoint for interaction in your IRIS app to interact with anything you've connected to HAOS.  Wanna restart an IRIS production with a flick of a physical switch or Assisted AI? You can do it with this add-on, or your own, right alongside the home automation hackers.

"Haul" a Portable Registry for Airgapped IrisClusters

Rancher Government Hauler streamlines deploying and maintaining InterSystems container workloads in air-gapped environments by simplifying how you package and move required assets. It treats container images, Helm charts, and other files as content and collections, letting you fetch, store, and distribute them declaratively or via CLI — without changing your existing workflows.   Meaning your charts and what have yous, can have conditionals on your pull locations in Helm values, etc. 

If you have been tracking how HealthShare is being deployed via IPM Packages, you can certainly appreciate the adoption of OCI compliance storage for the packages themselves using ORAS... which is core to the Hauler solution.

I am brand new to using AI. I downloaded some medical visit progress notes from my Patient Portal. I extracted text from PDF files. I found a YouTube video that showed how to extract metadata using an OpenAI query / prompt such as this one:

ollama-ai-iris/data/prompts/medical_progress_notes_prompt.txt at main · oliverwilms/ollama-ai-iris
 

I combined @Rodolfo Pscheidt Jr https://github.com/RodolfoPscheidtJr/ollama-ai-iris with some files from @Guillaume Rongier https://openexchange.intersystems.com/package/iris-rag-demo.

I attempted to run

python3 query_data.

I'm trying to deploy a python/flask application to an Iris4Health container (2025.1) but running into some issues with packages and where they should be installed. 

First one if the flask packages themselves. I tried to installed the flask packages into the python external language server virtual environment, but even doing this when configuring the WSGI application in the web application section, it would complain about not having a WSGI framework. Once I created a custom container and added the flask package at the OS level, I was able to configure the web application without it complaining.

I combined @Rodolfo Pscheidt https://github.com/RodolfoPscheidtJr/ollama-ai-iris with some files from @Guillaume Rongier https://openexchange.intersystems.com/package/iris-rag-demo.

My own project is https://github.com/oliverwilms/ollama-ai-iris

I can run load_data.py and it connects to IRIS (same container).

When I try to run query_data.py https://github.com/oliverwilms/ollama-ai-iris/blob/main/query_data.py , it cannot connect to ollama:

ConnectionError: Failed to connect to Ollama. Please check that Ollama is downloaded, running and accessible.

I wonder if I need to add in query_data.

If you look at the values.yaml of the IKO's Helm chart you'll find:

useIrisFsGroup: false 

Let's break down what it is and in what situations you may want to set it to true.

FsGroup refers to the file system group.

By default, Kubernetes volumes are owned by root, but we need IRIS to own its files (IRIS in containers is installed under irisowner user). To get around this we employ one of two methods:

1) initContainers

The initContainers run before app containers (like IRIS) in a pod. They generally set up the environment for the application and then run to completion/terminate.