I spent a lot of time practicing in labs that I built from Kubernetes the Hard Way and kubeadm. Using kubeadm is nice for a lot of things, but the experience building your own clusters helps out a lot during the troubleshooting sections of the exam.
Being able to create objects such as Deployments, Pods, etc using YAML is very helpful, but being able to generate YAML by using "kubectl get [object] -o yaml" can be a time saver. I tried to remember how to manually create every objects via YAML by memory and it came in helpful.
For documentation, I used docs.kubernetes.io and "kubectl explain". Both have their advantages depending on the situation. You should be familar with the Kubernetes documentation so when you pull it up, you know where to go. Usually you can search what you're looking for and it will be the first hit. Then you can scroll to the section you need. It's probably a bad sign if you're hitting a documentation page for the first time during the same. I did this twice :(
Definitely read the handbook and tips PDFs that are provided when you register for the exam. They will describe the clusters you'll be working with and you can get some insight on what problems you'll face just by looking at the clusters.
Lastly, during my studying homestretch, I tried to make it as much like the exam as possible. You're only allowed to have one terminal, one app (Chrome) and two tabs opened (one for the test and one for docs.kubernetes.io). I tried to get used to just having this setup so I wouldn't accidentally open more tabs and go somewhere I shouldn't.