How to connect Schrodinger to KNIME Business Hub?

yash_vyas · April 8, 2026, 2:12pm

Hi all,

I need to connect Schrödinger software to KNIME Business Hub. I already have a workflow running on KNIME 5.5.3 that includes Schrödinger nodes, and I have configured the licenses in the Preferences tab of the KNIME Analytics Platform.

I would also like to deploy this workflow to KNIME Business Hub so that it can use these licenses. However, setting up Schrödinger requires several steps, such as modifying the schrodinger.hosts file and configuring settings via the Preferences tab in the KNIME Analytics Platform GUI. I am unsure whether these steps can be performed on KNIME Business Hub, since the GUI is part of the executor, meaning there is no graphical interface available. I also do not know how to access the Preferences configuration from the command line.

My KNIME Business Hub is hosted on a Linux server. I understand that I will need to create a new executor using a custom Docker image that includes Schrödinger, and assign it the appropriate resources.

Could anyone help with a template or guide for achieving this? The goal is for users to be able to upload a file to the workflow and receive an output, with Schrödinger running in the background on the server.

Thanks in advance!

mjfratoni89 · April 13, 2026, 4:54pm

Hi @yash_vyas,

To set up Schrödinger on Business Hub, you’ll need the following:

Install KNIME Business Hub

Create Schrödinger license server
For the KNIME Business Hub, it is required that the customer has a Schrödinger server license. Because our executor containers are ephemeral, a static license will not work. They need to communicate with the license server to grab thread tokens for execution.
1. If doing a single-node install of BH, you can place on the same machine no problem
2. If doing a multi-node install of BH, you will need to evaluate where to put it. As long as all pods running Schrödinger have internet access to said machine, it is fine.
Create a Execution context with the following:
1. Schrödinger extension installed
  For the custom image, I’ve attached a template you can work off of (it’s for 5.1, so adjust where necessary) as well as the related required linux packages (packages.txt)
  
  To upload, follow the documentation to push image to the Hub registry and then create and execution context with that image:
  Docker Execution Images
  Execution Resouces
2. Persistent Volume Claim attached
  create a Persistent Volume Claim in the cluster and add Schrödinger software to it.
  - You can find the relevant download at the Schrödinger Downloads page
  - Make sure you download the correct version, see KNIME/Schrödinger compatibility matrix
3. Schrödinger software installed on the persistent volume
Connect execution context install to the Schrödinger license server

dockerfile.txt (1.1 KB)

packages.txt (479 Bytes)

mjfratoni89 · April 13, 2026, 5:09pm

For creating the PVC and mount to the Execution Context:

Create execution-additional-mount-pvc.yaml

apiVersion: v1
kind: PersistentVolumeClaim
metadata:
  name: additional-execution-mount-pvc
  namespace: hub-execution
spec:
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 50Gi

and run

kubectl apply -n hub-execution -f ./execution-additional-mount-pvc.yaml

Create a temporary pod for copying data onto volume:
Create mount-copy-pod.yaml

apiVersion: v1
kind: Pod
metadata:
  name: copy-to-pvc
  namespace: hub-execution
  labels:
    sidecar.istio.io/inject: "false"
spec:
  containers:
  - name: alpine
    image: alpine:latest
    command: ['sleep', 'infinity']
    volumeMounts:
    - name: additional-execution-mount-pvc
      mountPath: /data
  volumes:
  - name: additional-execution-mount-pvc
    persistentVolumeClaim:
      claimName: additional-execution-mount-pvc

and run

kubectl apply -n hub-execution -f ./mount-copy-pod.yaml

Copy the Schrödinger linux install file (.tar) to the persistent volume

kubectl cp -n hub-execution <path-to-tar-file> copy-to-pvc:/data

Delete attached temporary pod

kubectl -n hub-execution delete -f ./mount-copy-pod.yaml

Update relevant execution context to attach volume

Get list of Execution Contexts with

kubectl -n hub-execution get executioncontexts.hub.knime.com

Output: list of ECs - use the one with the Schrödinger nodes installed for in the next step

Retrieve JSON object for EC with (this command makes an ec.yaml file)

kubectl -n hub-execution get ``executioncontext.hub.knime.com/``<ec-id> -o yaml > ec.yaml

Modify the ec.yaml file to add the volume (add lines 9-15 under executor):

apiVersion: hub.knime.com/v1alpha1
kind: ExecutionContext
metadata:
  ...
spec:
  ...
  executor:
    ...
    volumeMounts:
    - mountPath: /external-tools
      name: additional-execution-mount-pvc
    volumes:
    - name: additional-execution-mount-pvc
      persistentVolumeClaim:
        claimName: additional-execution-mount-pvc
status:
  ...

and apply the changes to the EC pod

kubectl -n hub-execution apply -f ec.yaml

To install the Schrödinger software within the mount:

Using OpenLens, you can shell into the EC pod to run the Schrödinger install. Once inside you will need to unpack the .tar file and install the software. To test if this worked correctly, run a df command. You should see a drive available called /external-tools - move to it using cd /external-tools. Inside you will see the installation .tar file

Unpack the .tar file using

tar -xvf Schrodinger_Suites_<version>_Linux-x86_64.tar

This will create the Schrodinger_Suites_<version>_Linux-x86_64 directory.

Enter this directory and run the install script using sh ./INSTALL. This will start the installation manager

***Notice we are installing the Schrödinger software a second time. The first time was outside the cluster to use the license server. This install is on the pod level for the workflows to reference.

Change the directory value to /external-tools/schrodinger2023-3

Set the scratch directory to be /tmp

Notice there is a KNIME AP install bundled here. It shouldn’t affect anything but you can remove it

Once the install is complete, check to make sure the SCHRODINGER environment variable is set with

echo $SCHRODINGER

It should point to /external-tools/Schrodinger2023-3

***Now Schrödinger software is installed on the persistent drive! No additional configuration is needed as KNIME will use the SCHRODINGER environment variable to communicate.

Connect the pod Schrödinger install to the Schrödinger license server

Last step is to make sure the Pod can reach out to the license server to authenticate and grab tokens for execution.

Run the following command to connect the pod to the license server (run this on the pod level)

$SCHRODINGER/licadmin INSTALL -c port@server

and were generated when you set up the license server.