Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Evaluate volume expansion #174

Closed
wants to merge 2 commits into from
Closed

Evaluate volume expansion #174

wants to merge 2 commits into from

Conversation

solsson
Copy link
Contributor

@solsson solsson commented May 1, 2018

@solsson solsson added this to the 4.1 milestone May 1, 2018
@clintfred
Copy link

@solsson Have you been able to test this on GKE? I see the volumeClaimTemplate starts with 10Gi instead of 200. Can you characterize the impact of the resize on the Kafka cluster?

Slightly off topic, but perhaps related - in my development cluster, we sometimes end up wanting to completely start with a new cluster (zookeeper, brokers, registry), but retain the data. I have been running delete -f on the directories in reverse order, but after reading #167 and #154 maybe I should have just used apply -f and then deleted the old pods to pick up the new configmaps? It also looks like there is more dynamic config coming in 4.x?

@solsson
Copy link
Contributor Author

solsson commented May 16, 2018

@clintfred

Have you been able to test this on GKE? I see the volumeClaimTemplate starts with 10Gi instead of 200. Can you characterize the impact of the resize on the Kafka cluster?

I only created a new test cluster, but never got round to trying the expansion. If it works we'd create new clusters at 10Gi because that's the minimum PD size (SSDs can be smaller) in GKE. We can never really estimate data volumes at cluster creation. I'll post here if/when we've actually evaluated anything :)

I'll post in #154 regarding cluster replacement.

@clintfred
Copy link

In Kubernetes 1.11 release announcement

Support for online resizing of Persistent Volumes has been introduced as an alpha feature. This enables users to increase the size of PVs without having to terminate pods and unmount volume first. The user will update the PVC to request a new size and kubelet will resize the file system for the PVC.

@solsson
Copy link
Contributor Author

solsson commented Jun 28, 2018

And it looks like offline resize could actually make it to non-alpha GKE clusters in 1.11. Release notes say "Resizing PersistentVolumes after pod restart is now considered beta." with a reference to kubernetes/enhancements#284.

Kafka clusters should be operational with one broker taken down for maintenance, so if resize is reasonably quick there is no reason to wait for online resize. I'll try as soon as we get our hands on a 1.11 cluster. We've failed to do resize in 1.9.

@solsson solsson modified the milestones: 4.1, 5.0 - Java 11 Nov 28, 2018
@solsson
Copy link
Contributor Author

solsson commented Nov 28, 2018

#191 includes the volume expansion setting, but not the volume size change in Kafka. I don't think we should have that until volume expansion is mainstream. Or maybe we do, with the same rationale as d088b25?

@solsson
Copy link
Contributor Author

solsson commented Nov 28, 2018

Actually #191 does include the changes to default volume size. Ok then.

@solsson solsson closed this Nov 28, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants