cancel
Showing results for 
Search instead for 
Did you mean: 
Data Engineering
Join discussions on data engineering best practices, architectures, and optimization strategies within the Databricks Community. Exchange insights and solutions with fellow data engineers.
cancel
Showing results for 
Search instead for 
Did you mean: 

Cluster Configuration Best Practices

Sandy21
New Contributor III

I have a cluster with the configuration of 400 GB RAM, 160 Cores.

Which of the following would be the ideal configuration to use in case of one or more VM failures?

Cluster A: Total RAM 400 GB

      Total Cores 160

  Total VMs: 1

  400 GB/Exec & 160 cores/Exec

Cluster B: Total RAM 400 GB

      Total Cores 160

  Total VMs: 2

  200 GB/Exec & 80 cores/Exec

Cluster C: Total RAM 400 GB

      Total Cores 160

  Total VMs: 4

  100 GB/Exec & 40 cores/Exec

Cluster 😧 Total RAM 400 GB

      Total Cores 160

  Total VMs: 8

  50 GB/Exec & 20 cores/Exec

Cluster E: Total RAM 400 GB

      Total Cores 160

  Total VMs: 16

  25 GB/Exec & 10 cores/Exec

1 ACCEPTED SOLUTION

Accepted Solutions

karthik_p
Esteemed Contributor

@Santhosh Raj​ can you please confirm cluster sizes you are taking are related to driver and worker node. how much you want to allocate to Driver and Worker?

once we are sure about type of driver and worker we would like to pick, we need to enable autoscaling to have workers to handle your load. best practices wise you need to

  1. Initially check data size
  2. how do you want to run (all purpose or job cluster)
  3. Go from small cluster size and Increase based on performance that you are expecting

View solution in original post

1 REPLY 1

karthik_p
Esteemed Contributor

@Santhosh Raj​ can you please confirm cluster sizes you are taking are related to driver and worker node. how much you want to allocate to Driver and Worker?

once we are sure about type of driver and worker we would like to pick, we need to enable autoscaling to have workers to handle your load. best practices wise you need to

  1. Initially check data size
  2. how do you want to run (all purpose or job cluster)
  3. Go from small cluster size and Increase based on performance that you are expecting
Join 100K+ Data Experts: Register Now & Grow with Us!

Excited to expand your horizons with us? Click here to Register and begin your journey to success!

Already a member? Login and join your local regional user group! If there isn’t one near you, fill out this form and we’ll create one for you to join!