[e2e] Test against session job exception fetching error

Author: gyfora

e2e-tests/test_application_operations.sh

@@ -63,6 +63,7 @@ jm_pod_name=$(get_jm_pod_name $CLUSTER_ID)
 # Check the new JobManager recovering from latest successful checkpoint
 wait_for_logs $jm_pod_name "Restoring job $job_id from Checkpoint" ${TIMEOUT} || exit 1
 wait_for_logs $jm_pod_name "Completed checkpoint [0-9]+ for job" ${TIMEOUT} || exit 1
+verify_not_in_operator_log $jm_pod_name "Failed to fetch job exceptions" ${TIMEOUT} || exit 1
 wait_for_status flinkdep/flink-example-statemachine '.status.jobManagerDeploymentStatus' READY ${TIMEOUT} || exit 1
 wait_for_status flinkdep/flink-example-statemachine '.status.jobStatus.state' RUNNING ${TIMEOUT} || exit 1
 assert_available_slots 1 $CLUSTER_ID

e2e-tests/test_sessionjob_operations.sh

@@ -96,5 +96,6 @@ retry_times 5 30 "kubectl apply -f $APPLICATION_YAML" || exit 1
 
 wait_for_jobmanager_running $CLUSTER_ID $TIMEOUT
 wait_for_logs $jm_pod_name "Completed checkpoint [0-9]+ for job" ${TIMEOUT} || exit 1
+verify_not_in_operator_log $jm_pod_name "Failed to fetch job exceptions" ${TIMEOUT} || exit 1
 wait_for_status $SESSION_CLUSTER_IDENTIFIER '.status.jobManagerDeploymentStatus' READY ${TIMEOUT} || exit 1
 wait_for_status $SESSION_JOB_IDENTIFIER '.status.jobStatus.state' RUNNING ${TIMEOUT} || exit 1

e2e-tests/utils.sh

@@ -56,6 +56,19 @@ function wait_for_operator_logs {
   exit 1
 }
 
+function verify_not_in_operator_log {
+  local operator_pod_name=$1
+  local regex_to_check=$2
+  local timeout=$3
+  operator_pod_namespace=$(get_operator_pod_namespace)
+
+  if kubectl logs "${operator_pod_name}" -c flink-kubernetes-operator -n "${operator_pod_namespace}" | grep -E "${regex_to_check}" >/dev/null; then
+    echo "Found \"$2\" in operator logs."
+    return 1
+  fi
+  return 0
+}
+
 function wait_for_status {
   local resource=$1
   local status_path=$2