Hi,
Just wanted to know from the expert what are the Prons and Cons when DSE Inside Kubernetes Vs Outside Kubernetes?
Hey part I am expecting
I am using DSE for a long time but now we are moving our infra to K8 but not abe to final call whether we should keep DSE inside K8 or outside K8
Looking forward to hearing from the expert.
Thanks
1. Performance: The K8ssandra group recently ran a suite of benchmarks comparing VM-based clusters with those running on Kubernetes. See https://k8ssandra.io/blog/articles/k8ssandra-performance-benchmarks-on-cloud-managed-kubernetes/
2. Scale up and down: When running in a k8s environment the operator handles all tasks associated with adding and removing nodes. Human operators simply indicate a number of nodes and it handles the rest. This includes support for stopping the cluster and scaling compute resources to 0 while persisting data.
3. Throughput: See the blog post in the Performance section for both throughput and latency examples.
4. Management: I've personally managed hundreds of nodes on bare-metal and VM-backed instances. Management of DSE on Kubernetes is a breeze with the tedious tasks of managing the cluster handled via the operator. Additionally better integration with centralized services like monitoring and log collection really makes this a winning combo.
Hi @ranjeet_ranjee, could you please update your original question with additional specifics by elaborating on it a little bit more to give the context, what you're trying to achieve, etc., for the experts to be able to help assist you with this? Thanks!
You could read about DataStax Enterprise (aka DSE) to understand about it a little more in the meantime.
The real reason for deploying any application or database on Kubernetes is the simplified deployment, management and day-to-day operations.
Kubernetes is an orchestration/automation platform. When you deploy DSE on Kubernetes, most of the operations such as scaling up/down is a lot simpler. In addition, Kubernetes automatically provisions replacement nodes when DSE pods go down for example when a Kubernetes node experiences a hardware failure.
The biggest drawback is that you need to invest in learning Kubernetes if you want to deploy apps in a Kubernetes cluster. This is another layer of complexity you need to be knowledgeable on and takes a while to build experience.
Note that you will still need a valid subscription to use DSE even when deployed in a Kubernetes cluster. For more information, see the DataStax Kubernetes Operator for Apache Cassandra. Cheers!
7 People are following this question.
