We use our own tools for spot clusters, but here is a gist for the userdata we use for the tservers in case it’s helpful. The master userdata is nearly identical. I found the only libs necessary to amend in Amazon Linux was the newer libsasl that Kudu requires. With EMR you’ll just need a way to communicate master address(es) to the tservers.
https://gist.github.com/cresny/d47fad80fd37ec41bde8871ac434d915 From: Dong Jiang <[email protected]<mailto:[email protected]>> Reply-To: "[email protected]<mailto:[email protected]>" <[email protected]<mailto:[email protected]>> Date: Thursday, July 20, 2017 at 8:11 PM To: "[email protected]<mailto:[email protected]>" <[email protected]<mailto:[email protected]>> Subject: Kudu bootstrap on EMR? Hi, I am wondering if anyone in the Kudu community has tried to install Kudu on AWS EMR via bootstrap action? If so, do you have a bootstrap script to share? Thanks, Dong
