This repository has been archived by the owner on Apr 4, 2023. It is now read-only.
Cassandra E2E tests fail on a multi-node cluster because we use hostpath volumes #351
Labels
Comments
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
In the E2E tests, we configure Cassandra nodes with hostpath volumes.
These work on minikube because it only creates a single node.
On a multi-node cluster such as created by kubeadm-dind-cluster, Cassandra Pods may get moved between nodes and in this case new (empty) hostpath volumes are created on each node and data appears to be lost.
Instead we could use persistent local storage PVs which will ensure that pods are only scheduled to nodes with the PV.
#330 has some example code for creating persistent local volumes for E2E tests.
/kind bug
The text was updated successfully, but these errors were encountered: