The cpu and memory usage are significantly higher

[tomdong2010]

Bug Description

when restart , the kafka cost cpu and memory is so big ,make my service stuck . And installing kafka is also to long time

Steps to reproduce

No response

Expected behavior

No response

Strimzi version

0.36.0

Kubernetes version

3.5

Installation method

yaml

Infrastructure

No response

Configuration files and logs

No response

Additional context

No response

[ppatierno]

Hi,
there is pretty much nothing useful to us in the description. We would need more information. At least, the Kafka CR you are using, the environment you are running, the cluster operator log.

[tomdong2010]

https://github.com/tomdong2010/operator_log/blob/main/strimzi-cluster-operator-7b4f8f8496-h8vtr_strimzi-cluster-operator.log you can see ， this installation process has been going on for 10 minutes my cpu average is 4 and memory about average 5Gi if i use native installation ,this situation is not appear , i tried many times

[ppatierno]

Where is the Kafka CR? Is it minikube? Kubernetes? You should describe the ZK pods, I think you have resources issues and they are not schedulable somehow?

[tomdong2010]

in k8s , but i also tried to use native installation , no such problem

[ppatierno]

what does native installation means?
I also asked for other information see my comments above.

[tomdong2010]

use helm to install zookeeper and kafka, there is not problem about cpu and memory. and also smooth, this is only my feeling ,maybe my problem . that`s all

[ppatierno]

we cannot compare apples and pears ... I would like to help but you are still not providing what I need. See my comments above.

[scholzj]

The operator log you shared suggests that the ZooKeeper pods were not ready. But the screenshots shows them as ready. So they do not seem to match with each other.

[tomdong2010]

https://github.com/tomdong2010/operator_log/blob/main/zkpod.png this is pod situation , normal https://github.com/tomdong2010/operator_log/blob/main/pods.png

[ppatierno]

Thanks Jakub, that is exactly what I was going to say.

[ppatierno]

https://github.com/tomdong2010/operator_log/blob/main/zkpod.png this is pod situation , normal https://github.com/tomdong2010/operator_log/blob/main/pods.png

So I don't understand now, is it working?

[tomdong2010]

I found the operator restart a time just now

[tomdong2010]

yep , i used it ok , but But the cpu and memory usage are more obvious . always appear this fail https://github.com/tomdong2010/operator_log/blob/main/fail.png

[scholzj]

I think we should probably change this to discussion since there does not seem to be anything suggesting any bug so far and you are not providing the information needed to really do anything about this and your screenshots actually suggest everything works fine one time and doesn't work next time which sounds like an environment issue.

If you think it is consuming too many resources, you should share the information you were asked in the beginning:

• The Kafka custom resource
• Full logs from all components (operator, ZooKeeper nodes, Kafka nodes etc.)
• Explain what you environment is (Strimzi works only on Kubernetes -> so the environment is always Kubernetes. More interesting are things such as Kubernetes distribution used, platform you use, version etc.)
• And you should also explain what does it mean that it is consuming too many resources. Which component, how do you measure it, share the charts showing the resource consumption together with the corresponding logs etc.

[tomdong2010]

ok ,but above image is a error about Readiness probe failed , this issue always appeared . maybe this is reason. change this to discussion

[ppatierno]

The thing is that for me and Jakub it's really difficult to help if you continue to share just screenshot.
Jakub provided you a more clear list of what we need to be helpful. Without them there is nothing we can do.
I would just go through that list and provide everything we asked for.