[FLINK-23418][e2e] Increase the timeout to make kubernetes applicatio…

…n ha test more stable

This closes apache#16602.

flink-end-to-end-tests/test-scripts/common_kubernetes.sh

@@ -171,22 +171,22 @@ function wait_num_checkpoints {
 function wait_for_logs {
   local jm_pod_name=$1
   local successful_response_regex=$2
+  local timeout=${3:-30}
 
   echo "Waiting for jobmanager pod ${jm_pod_name} ready."
-  kubectl wait --for=condition=Ready --timeout=30s pod/$jm_pod_name || exit 1
+  kubectl wait --for=condition=Ready --timeout=${timeout}s pod/$jm_pod_name || exit 1
 
-  # wait at most 30 seconds until the log shows up
-  local TIMEOUT=30
+  # wait or timeout until the log shows up
   echo "Waiting for log \"$2\"..."
-  for i in $(seq 1 ${TIMEOUT}); do
+  for i in $(seq 1 ${timeout}); do
     if check_logs_output $jm_pod_name $successful_response_regex; then
       echo "Log \"$2\" shows up."
       return
     fi
 
     sleep 1
   done
-  echo "Log $2 does not show up within a timeout of ${TIMEOUT} sec"
+  echo "Log $2 does not show up within a timeout of ${timeout} sec"
   exit 1
 }
 

flink-end-to-end-tests/test-scripts/test_kubernetes_application_ha.sh

@@ -74,7 +74,7 @@ job_id=$(kubectl logs $jm_pod_name | grep -E -o 'Job [a-z0-9]+ is submitted' | a
 kubectl exec $jm_pod_name -- /bin/sh -c "kill 1"
 
 # Check the new JobManager recovering from latest successful checkpoint
-wait_for_logs $jm_pod_name "Restoring job $job_id from Checkpoint"
+wait_for_logs $jm_pod_name "Restoring job $job_id from Checkpoint" 120
 wait_num_checkpoints $jm_pod_name 1
 
 "$FLINK_DIR"/bin/flink cancel -t kubernetes-application -Dkubernetes.cluster-id=${CLUSTER_ID} $job_id