Mismatched datadir volumeMount causing zookeeper dataloss after pod recreation #2812
Open
1 of 3 tasks
Labels
kind/k8s
Related to the Kubernetes application
needs-triage
This wasn't investigate by the repo's owners yet
Category:
Kubernetes apps
Type:
We are using the GKE click-to-deploy feature to deploy Kafka in our cluster, but after the underlying nodes are replaced, Kafka fails to start properly and throws the following error:
After our investigation, we found that there might be an issue with the Zookeeper configuration in the chart. The
datadir
volumeMount configuration for Zookeeper is inconsistent with the value of theZK_DATA_DIR
environment variable, causing data loss in Zookeeper after pod migration.In
k8s/kafka/chart/kafka/templates/zk-statefulset.yaml
,ZK_DATA_DIR
is set to/data
butvolumeMounts
is configured as:when we checked the
/opt/zookeeper
folder inside the pod, we found it was empty.The text was updated successfully, but these errors were encountered: