I was trying to install SQL Server 2017 on my latest formatted machine and was stuck at below screen which says – Offline Installation of Microsoft Machine Learning Server Components. In this blog, we would learn about how to move forward in the Installation when it is stuck at Offline Installation of Microsoft Machine Learning Server Components.
First let’s understand the cause the screen because when I was installing SQL with my client, I didn’t see this screen at all.
It looks like If we select Machine Learning, the setup would connect to the internet and download few components. If it is not able to reach to the internet, then it would provide above the screen.
I looked at Detail.txt and then RSetup.log and found below in RSetup.log.
2018-11-26T03:48:50 INFO Command invoked: C:\Program Files\Microsoft SQL Server\140\Setup Bootstrap\SQL2017\x64\RSetup.exe /checkurl /component SRS /version 184.108.40.206 /language 1033 /logfile C:\Program Files\Microsoft SQL Server\140\Setup Bootstrap\Log\20181125_193933\RSetup.log
2018-11-26T03:48:50 INFO RSetup.exe version: 220.127.116.11
2018-11-26T03:48:50 INFO Validating URL: https://rserverdistribution.azureedge.net/production/SRS/18.104.22.168/3013/1033/33f8d690e29044b4b6ff3d3989c8756d/SRS_22.214.171.124_1033.cab
2018-11-26T03:49:11 WARN Error making request: Unable to connect to the remote server
2018-11-26T03:49:11 WARN Error validating download
2018-11-26T03:49:11 INFO Exiting with code 851507
All it says that it is not able to reach to the internet and download the packages. This depends on what features were selected during the feature selection screen. A picture speaks a thousand words and below two images would explain what I mean.
As explained in the message, we need to download ALL the files from ALL the link provided below and then keep them in some folder. Here are the links for your ready reference.
Here are the URLs which come when “R” is selected.
Here are the URLs which would come when “Python” is selected.
Hope this would help someone and save time which I spent in finding the behavior/solution.
Reference: Pinal Dave (https://blog.sqlauthority.com)