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

#!/usr/bin/env bash
################################################################################
# Licensed to the Apache Software Foundation (ASF) under one
# or more contributor license agreements.  See the NOTICE file
# distributed with this work for additional information
# regarding copyright ownership.  The ASF licenses this file
# to you under the Apache License, Version 2.0 (the
# "License"); you may not use this file except in compliance
# with the License.  You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
################################################################################
set -o pipefail

source "$(dirname "$0")"/common.sh
source "$(dirname "$0")"/common_docker.sh
source "$(dirname "$0")"/common_artifact_download_cacher.sh

FLINK_TARBALL_DIR=$TEST_DATA_DIR
FLINK_TARBALL=flink.tar.gz
FLINK_DIRNAME=$(basename $FLINK_DIR)

MAX_RETRY_SECONDS=120
CLUSTER_SETUP_RETRIES=3
IMAGE_BUILD_RETRIES=5

echo "Flink Tarball directory $FLINK_TARBALL_DIR"
echo "Flink tarball filename $FLINK_TARBALL"
echo "Flink distribution directory name $FLINK_DIRNAME"
echo "End-to-end directory $END_TO_END_DIR"

start_time=$(date +%s)

# make sure we stop our cluster at the end
function cluster_shutdown {
  if [ $TRAPPED_EXIT_CODE != 0 ];then
      debug_copy_and_show_logs
  fi
  docker-compose -f $END_TO_END_DIR/test-scripts/docker-hadoop-secure-cluster/docker-compose.yml down
  rm $FLINK_TARBALL_DIR/$FLINK_TARBALL
}
on_exit cluster_shutdown

function start_hadoop_cluster() {
    echo "Starting Hadoop cluster"
    docker-compose -f $END_TO_END_DIR/test-scripts/docker-hadoop-secure-cluster/docker-compose.yml up -d

    # wait for kerberos to be set up
    start_time=$(date +%s)
    until docker logs master 2>&1 | grep -q "Finished master initialization"; do
        current_time=$(date +%s)
        time_diff=$((current_time - start_time))

        if [ $time_diff -ge $MAX_RETRY_SECONDS ]; then
            return 1
        else
            echo "Waiting for hadoop cluster to come up. We have been trying for $time_diff seconds, retrying ..."
            sleep 5
        fi
    done

    # perform health checks
    containers_health_check "master" "worker1" "worker2" "kdc"

    # try and see if NodeManagers are up, otherwise the Flink job will not have enough resources
    # to run
    nm_running="0"
    start_time=$(date +%s)
    while [ "$nm_running" -lt "2" ]; do
        current_time=$(date +%s)
        time_diff=$((current_time - start_time))

        if [ $time_diff -ge $MAX_RETRY_SECONDS ]; then
            return 1
        else
            echo "We only have $nm_running NodeManagers up. We have been trying for $time_diff seconds, retrying ..."
            sleep 1
        fi

        docker exec master bash -c "kinit -kt /home/hadoop-user/hadoop-user.keytab hadoop-user"
        nm_running=`docker exec master bash -c "yarn node -list" | grep RUNNING | wc -l`
        docker exec master bash -c "kdestroy"
    done

    echo "We now have $nm_running NodeManagers up."

    return 0
}

function build_image() {
    echo "Predownloading Hadoop tarball"
    cache_path=$(get_artifact "http://archive.apache.org/dist/hadoop/common/hadoop-2.8.4/hadoop-2.8.4.tar.gz")
    ln "$cache_path" "$END_TO_END_DIR/test-scripts/docker-hadoop-secure-cluster/hadoop-2.8.4.tar.gz"

    echo "Building Hadoop Docker container"
    docker build --build-arg HADOOP_VERSION=2.8.4 \
        -f $END_TO_END_DIR/test-scripts/docker-hadoop-secure-cluster/Dockerfile \
        -t flink/docker-hadoop-secure-cluster:latest \
        $END_TO_END_DIR/test-scripts/docker-hadoop-secure-cluster/
}

function start_hadoop_cluster_and_prepare_flink() {
    if ! retry_times $IMAGE_BUILD_RETRIES 2 build_image; then
        echo "ERROR: Could not build hadoop image. Aborting..."
        exit 1
    fi
    if ! retry_times $CLUSTER_SETUP_RETRIES 0 start_hadoop_cluster; then
        echo "ERROR: Could not start hadoop cluster. Aborting..."
        exit 1
    fi

    mkdir -p $FLINK_TARBALL_DIR
    tar czf $FLINK_TARBALL_DIR/$FLINK_TARBALL -C $(dirname $FLINK_DIR) .

    docker cp $FLINK_TARBALL_DIR/$FLINK_TARBALL master:/home/hadoop-user/

    # now, at least the container is ready
    docker exec master bash -c "tar xzf /home/hadoop-user/$FLINK_TARBALL --directory /home/hadoop-user/"

    # minimal Flink config, bebe
    FLINK_CONFIG=$(cat << END
security.kerberos.login.keytab: /home/hadoop-user/hadoop-user.keytab
security.kerberos.login.principal: hadoop-user
slot.request.timeout: 120000
END
)
    docker exec master bash -c "echo \"$FLINK_CONFIG\" > /home/hadoop-user/$FLINK_DIRNAME/conf/flink-conf.yaml"

    echo "Flink config:"
    docker exec master bash -c "cat /home/hadoop-user/$FLINK_DIRNAME/conf/flink-conf.yaml"
}

function debug_copy_and_show_logs {
    echo "Debugging failed YARN Docker test:"
    echo -e "\nCurrently running containers"
    docker ps

    echo -e "\n\nCurrently running JVMs"
    jps -v

    local log_directory="$TEST_DATA_DIR/logs"
    local yarn_docker_containers="master $(docker ps --format '{{.Names}}' | grep worker)"

    extract_hadoop_logs ${log_directory} ${yarn_docker_containers}
    print_logs ${log_directory}

    echo -e "\n\n ==== Flink logs ===="
    docker exec master bash -c "kinit -kt /home/hadoop-user/hadoop-user.keytab hadoop-user"
    docker exec master bash -c "yarn application -list -appStates ALL"
    application_id=`docker exec master bash -c "yarn application -list -appStates ALL" | grep -i "Flink" | grep -i "cluster" | awk '{print \$1}'`

    echo -e "\n\nApplication ID: '$application_id'"
    docker exec master bash -c "yarn logs -applicationId $application_id"

    docker exec master bash -c "kdestroy"
}

function extract_hadoop_logs() {
    local parent_folder="$1"
    shift
    docker_container_aliases="$@"

    for docker_container_alias in $(echo ${docker_container_aliases}); do
        local target_container_log_folder="${parent_folder}/${docker_container_alias}"
        echo "Extracting ${docker_container_alias} Hadoop logs into ${target_container_log_folder}"
        mkdir -p "${target_container_log_folder}"
        docker cp "${docker_container_alias}:/var/log/hadoop/" "${target_container_log_folder}"

        local target_container_docker_log_file="${target_container_log_folder}/docker-${docker_container_alias}.log"
        echo "Extracting ${docker_container_alias} Docker logs into ${target_container_docker_log_file}"
        docker logs "${docker_container_alias}" > "${target_container_docker_log_file}"
    done
}

function print_logs() {
    local parent_folder="$1"

    ls -lisahR "${parent_folder}"
    find "${parent_folder}" -type f -exec echo -e "\n\nContent of {}:" \; -exec cat {} \;
}

# expects only one application to be running and waits until this one is in
# final state SUCCEEDED
function wait_for_single_yarn_application {

    docker exec master bash -c "kinit -kt /home/hadoop-user/hadoop-user.keytab hadoop-user"

    # find our application ID
    docker exec master bash -c "yarn application -list -appStates ALL"
    application_id=$(docker exec master bash -c "yarn application -list -appStates ALL" | grep "Flink Application" | awk '{print $1}')

    echo "Application ID: $application_id"

    # wait for the application to finish succesfully
    start_time=$(date +%s)
    application_state="UNDEFINED"
    while [[ $application_state != "FINISHED" ]]; do
        current_time=$(date +%s)
        time_diff=$((current_time - start_time))

        if [[ $time_diff -ge $MAX_RETRY_SECONDS ]]; then
            echo "Application $application_id is in state $application_state and we have waited too long, quitting..."
            exit 1
        else
            echo "Application $application_id is in state $application_state. We have been waiting for $time_diff seconds, looping ..."
            sleep 1
        fi

        application_state=$(docker exec master bash -c "yarn application -status $application_id" | grep "\sState" | sed 's/.*State : \(\w*\)/\1/')
    done

    final_application_state=$(docker exec master bash -c "yarn application -status $application_id" | grep "\sFinal-State" | sed 's/.*Final-State : \(\w*\)/\1/')

    echo "Final Application State: $final_application_state"

    if [[ $final_application_state != "SUCCEEDED" ]]; then
        echo "Running the Flink Application failed. 😞"
        exit 1
    fi

    docker exec master bash -c "kdestroy"
}

function get_output {
    docker exec master bash -c "kinit -kt /home/hadoop-user/hadoop-user.keytab hadoop-user"
        docker exec master bash -c "hdfs dfs -ls $1"
        OUTPUT=$(docker exec master bash -c "hdfs dfs -cat $1")
        docker exec master bash -c "kdestroy"
        echo "$OUTPUT"
}