Initial Configuration

When you first sign up for the InfinStor service, you need to configure your own AWS account to provide permissions for the InfinStor service to perform certain activities. These include:

  • Create an S3 bucket for storing mlflow artifacts
  • Allow the InfinStor service to create EC2 VMs for running your transforms
  • Allow the InfinStor service to start VMs on your behalf for running transforms, hosted jupyterlab instances, etc.
  • Direct S3 Events from your S3 buckets to the InfinStor service for creating InfinSnap and InfinSlice metadata (InfinStor Premium and higher only)

The configuration is performed to two consecutive steps.

  • Step 1: Enable InfinStor MLflow tracking service
  • Step 2: Enable InfinStor compute, i.e. starting VMs for processing transforms, etc.

This configuration is performed in two steps in order to give you the option of using mlfow tracking, but not hosted jupyterlab, transforms, etc.

Step 1: Enable InfinStor MLflow tracking service

InfinStor MLflow tracking service is enabled by downloading and running the enable-mlflow-tracking.sh script. The easiest way to do this is to run this script in AWS Cloudshell

Start the AWS CloudShell from the AWS Console, and then run the following command. You can conveniently copy using the icon in the top right of the code block below and paste into CmdShell:

/bin/rm -f ./enable-mlflow-tracking.sh && curl -s https://service.infinstor.com/tools/enable-mlflow-tracking.sh -o enable-mlflow-tracking.sh && bash ./enable-mlflow-tracking.sh && /bin/rm -f ./enable-mlflow-tracking.sh

Screencapture of AWS CloudShell running enable-mlflow-tracking.sh

Step 2: Enable InfinStor Storage

After running Step 1 above, i.e. enabling InfinStor MLflow tracking service, you can enable InfinStor Storage by downloading and running the enable-storage.sh script. The easiest way to do this is to run this script in AWS Cloudshell

Start the AWS CloudShell from the AWS Console, and then run the following command:

/bin/rm -f ./enable-storage.sh && curl -s https://service.infinstor.com/tools/enable-storage.sh -o enable-storage.sh && bash ./enable-storage.sh && /bin/rm -f ./enable-storage.sh

Step 3: Enable InfinStor Compute

After running Steps 1 and 2 above, i.e. enabling InfinStor MLflow tracking service and InfinStor Storage, you can enable InfinStor Compute service by downloading and running the enable-compute.sh script. The easiest way to do this is to run this script in AWS Cloudshell

Start the AWS CloudShell from the AWS Console, and then run the following command:

/bin/rm -f ./enable-compute.sh && curl -s https://service.infinstor.com/tools/enable-compute.sh -o enable-compute.sh && bash ./enable-compute.sh && /bin/rm -f ./enable-compute.sh

Screencapture of AWS CloudShell running enable-compute.sh

Note that in the above screencapture, enable-compute.sh asks whether you want to create a custom VPC for InfinStor, or use the default VPC.