Date
1 - 6 of 6
Need help setting UP Strimzi Kafka Cluster with Cruise Control
|
Ranjeet Ranjan
Hi Team,
Let me introduce myself Ranjeet Ranjan, CEO & Founder Adsizzler Media. We have recently launched one new Ad Tech Startup Named http://adosiz.com/ We wanted to install but need help setting UP Strimzi Kafka Cluster with Cruise Control for Auto Scale. I have downloaded helm from https://strimzi.io/downloads/. I followed the below steps to install. Step -1: helm install strimzi-kafka-operator strimzi-kafka-operator -n test But it only installed strimzi-cluster-operator-5f56c65c5d-4pphv Now please let know how Kafka Cluster pod will be created. Thanks Ranjeet Ranjan
|
|
|
|
Jakub Scholz
HI Ranjeet, The Helm Chart installs only the operator for managing the Kafka clusters. Next, you need to create the Kafka custom resource describing the Kafka cluster you want to have, its configuration etc. The docs (https://strimzi.io/documentation/) cover all the different aspects which you can configure. We have also all kind of examples on GitHub: https://github.com/strimzi/strimzi-kafka-operator/tree/main/examples Thanks & Regards Jakub On Tue, Feb 22, 2022 at 6:33 PM Ranjeet Ranjan <ranjeet@...> wrote:
|
|
|
|
Ranjeet Ranjan
Hi Jakub, Thank you for your response. DO we have a helm as well for creating Kafka Cluster? On Tue, Feb 22, 2022 at 11:08 PM Jakub Scholz <jakub@...> wrote:
|
|
|
|
Jakub Scholz
No, we do not have a Helm Chart for creating the Kafka custom resource. Thanks & Regards Jakub On Tue, Feb 22, 2022 at 6:40 PM Ranjeet Ranjan <ranjeet@...> wrote:
|
|
|
|
Ranjeet Ranjan
Hi Jakub, I have set up Strimzi Kafka Cluster and Operator. I have still one doubt that I did not answer, do I need to scale the Kafka cluster and data balance up/down manually or will it scale up/down based on the Kubernetes HPA? Your help will be much appreciated. On Tue, Feb 22, 2022 at 11:47 PM Jakub Scholz <jakub@...> wrote:
|
|
|
|
Jakub Scholz
You have to do it manually. Kafka is not something that is normally scaled up or down very often. It is a stateful application with partitions and their replicas having a fixed assignment to a specific broker. Scaling the cluster up and down means reassigning the partitions between the nodes and since the partitions often contain very large amounts of data, these data need to be transferred. And that has of course its cost (it often takes a long time, generates financial costs etc.). So usually, you would scale the cluster up for example as you onboard new applications or as your applications grow. But you would for example not scale it up for a few minutes because there seems to be some traffic peak. So normally it is more the result of some long term planning rather than some quick decision of the HPA. Thanks & Regards Jakub On Wed, Feb 23, 2022 at 9:48 PM Ranjeet Ranjan <ranjeet@...> wrote:
|
|
|