Known issues in Impala Virtual Warehouses

This topic describes the Impala Virtual Warehouse known issues for Cloudera Data Warehouse Private Cloud.

DWX-10382: (Affected version: 1.4.0) Unable to connect to the pre-upgrade Impala Virtual Warehouse using impala-shell on an Embedded Container Service (ECS) cluster

After upgrading from 1.3.2, 1.3.3, or 1.3.4 to 1.4.0, you may not be able to access the pre-upgrade Impala coordinators through its endpoints.

Restarting the coordinator pods can solve the issue. Alternatively, you can patch the leader elector image in Impala coordinators after upgrading CDW as follows. This is necessary because the old leader elector image runs deprecated code which is not compatible with the latest RKE versions:

1. Start a terminal session that can connect to your Kubernetes cluster using kubeconfig, for example.
2. Run the following script:

   #!/usr/bin/env bash
   set -x
   
   DWX_VERSION="1.4.0-b26"
   
   NAMESPACE=${NAMESPACE:?please set the NAMESPACE to update}
   
   LEADER_ELECTOR_IMAGE_PRE_UPGRADE=$(kubectl get sts coordinator -n $NAMESPACE \
     -o jsonpath={.spec.template.spec.containers[3].image})
   
   LEADER_ELECTOR_IMAGE_POST_UPGRADE=$(sed \
     "s/:.*/:$DWX_VERSION/" <<< $LEADER_ELECTOR_IMAGE_PRE_UPGRADE)
   
   kubectl patch sts coordinator -n $NAMESPACE --type='json' -p='[{"op": "replace", "path": "/spec/template/spec/containers/3/image", "value":"'$LEADER_ELECTOR_IMAGE_POST_UPGRADE'"}]'
   
   kubectl delete pod -l app=coordinator