Some keys are missing after autoscale on Kubernetes #60
Labels
bug
Something isn't working
help wanted
Extra attention is needed
investigation
olric
Distributed cache and in-memory key/value store
release-blocker
Milestone
Olric losts a relatively small number of keys after autoscaling in the Kubernetes environment.
Reproduce & observe:
1- Create a deployment with replicas: 1
2- Configure HPA and metrics server: https://kubernetes.io/docs/tasks/run-application/horizontal-pod-autoscale-walkthrough/
3- Assign an external IP to the deployment
4- Autoscale the deployment:
kubectl autoscale deployment olricd --cpu-percent=50 --min=1 --max=10
5- Run olric-load on your host system:
olric-load -a 10.98.198.52:3320 -c put -k 1000000 -s msgpack
6- It should run and scale up seamlessly.
7- Check all the keys:
olric-load -a 10.98.198.52:3320 -c get-k 1000000 -s msgpack
8- Some keys are missing:
[ERROR] Failed to call Get command for 209813: key not found
The text was updated successfully, but these errors were encountered: