You can use smaller instances with less ram than the VMs with higher ram, However there will be a trade off if there are lots of shuffle involve in the operation, because more small memory vms will increase the shuffling operation time
You can use smaller instances with less ram than the VMs with higher ram, However there will be a trade off if there are lots of shuffle involve in the operation, because more small memory vms will increase the shuffling operation time
You can also tune the JVM's GC parameters directly, if you mean the pauses are too long. Set "spark.executor.extraJavaOptions", but it does require knowing a thing or two about how to tune for what performance goal.
Welcome to Databricks Community: Lets learn, network and celebrate together
Join our fast-growing data practitioner and expert community of 80K+ members, ready to discover, help and collaborate together while making meaningful connections.