Python Package Installation on BDC

Shobhit Sinha 1

Hi Team,

We have setup azure sql bdc in azure environment. As per the Microsoft document in order to install packages in sql bdc using ADS pyspark kernel, there are two approaches:
1: Using sc._install(<package name>)
2: Using requirements.txt file.

Approach one is working fine in our setup but approach two is failing as kernel is getting dead while execution. Therefore please let us know if there are any configuration/pre requisites required to install packages using requirements.txt approach.

Microsoft doc url to install packages:
https://learn.microsoft.com/en-us/sql/big-data-cluster/spark-install-packages?view=sql-server-ver16

deherman-MSFT 33,141 Reputation points Microsoft Employee

2022-06-15T18:26:06.057+00:00

@Shobhit Sinha Thanks for your question. Could you give me more details on your environment and setup? What is the OS that you are running? Is this on an Azure VM? Have you tried using a python script to deploy a SQL Server big data cluster on Azure Kubernetes Service (AKS)?
Shobhit Sinha 1 Reputation point

2022-06-19T04:16:06.61+00:00

Please find the details below:

BDC is deployed on Azure VM with following configuration
OS: Ubuntu
Size: Standard DS12 v2 (4 vcpus, 28 GiB memory)
Shobhit Sinha 1 Reputation point

2022-06-19T04:16:53.657+00:00

For accessing BDC we have setup another Azure VM with following configuration
OS: Windows (Windows Server 2012 R2 Datacenter)
Size:- Standard DS12 v2 (4 vcpus, 28 GiB memory)
Tool: Azure Data Studio(Pyspark kernel)

We have not used python script to deploy a SQL Big Data cluster on Azure Kubernetes Service. For the deployment of BDC we have used this script on github repo.
[Link][1]