To answer one of the questions of specifying which nodes to run the executors that already has a locally pulled image, I'm not sure about spark drivers specifically, but I understand Kubernetes has node affinities and node selectors to allow users to specify which nodes to run pods on. So maybe that can solve the latency issue from having to pull a new image on a foreign node
Please advise. I am replicating the solution from this video on a Centos server which also acts as a docker env. Getting errors "none driver doesn't support minikube-env" & "cannot find image" when running docker-image-tool.sh. Please advise how to fix it?
A great explanation and has cleared a lot of my doubts on Kubernetes and how it works along with Spark. Thanks!
To answer one of the questions of specifying which nodes to run the executors that already has a locally pulled image, I'm not sure about spark drivers specifically, but I understand Kubernetes has node affinities and node selectors to allow users to specify which nodes to run pods on. So maybe that can solve the latency issue from having to pull a new image on a foreign node
Sweet - thank you gents
Excellent!
Please advise. I am replicating the solution from this video on a Centos server which also acts as a docker env. Getting errors "none driver doesn't support minikube-env" & "cannot find image" when running docker-image-tool.sh. Please advise how to fix it?
you got a image registry ?
480p on 2018...
can you share me this guild and github repository
Please share the code - it's been a while ;)
Excellent