Я пытаюсь развернуть задание Spark в Kubernetes, используя kubectl apply -f <config_file.yml> (после построения Docker образа на основе Dockerfile). Модуль успешно создается на K8s, а затем быстро останавливается со статусом Failed. Ничто в журналах не помогает понять, откуда возникла ошибка. Другие задания были успешно развернуты в кластере K8s с использованием того же файла Dockerfile и файла конфигурации.
Предполагается, что задание spark считывает данные из kafka topi c, анализирует их и выводит их на консоль.
Есть идеи, что может вызвать сбой работы?
Dockerfile, построенный с использованием docker build --rm -f "Dockerfile" xxxxxxxx:80/apache/myapp-test . && docker push xxxxxxxx:80/apache/myapp-test:
FROM xxxxxxxx:80/apache/spark:v2.4.4-gcs-prometheus
#USER root
ADD myapp.jar /jars
RUN adduser --no-create-home --system spark
RUN chown -R spark /prometheus /opt/spark
USER spark
config_file.yml:
apiVersion: "sparkoperator.k8s.io/v1beta2"
kind: SparkApplication
metadata:
name: myapp
namespace: spark
labels:
app: myapp-test
release: spark-2.4.4
spec:
type: Java
mode: cluster
image: "xxxxxxxx:80/apache/myapp-test"
imagePullPolicy: Always
mainClass: spark.jobs.app.streaming.Main
mainApplicationFile: "local:///jars/myapp.jar"
sparkVersion: "2.4.4"
restartPolicy:
type: OnFailure
onFailureRetries: 5
onFailureRetryInterval: 30
onSubmissionFailureRetries: 0
onSubmissionFailureRetryInterval: 0
driver:
cores: 1
memory: "1G"
labels:
version: 2.4.4
monitoring:
exposeDriverMetrics: true
exposeExecutorMetrics: true
prometheus:
jmxExporterJar: "/prometheus/jmx_prometheus_javaagent-0.11.0.jar"
port: 8090
imagePullSecrets:
- xxx
Журналы:
++ id -u
+ myuid=100
++ id -g
+ mygid=65533
+ set +e
++ getent passwd 100
+ uidentry='spark:x:100:65533:Linux User,,,:/home/spark:/sbin/nologin'
+ set -e
+ '[' -z 'spark:x:100:65533:Linux User,,,:/home/spark:/sbin/nologin' ']'
+ SPARK_K8S_CMD=driver
+ case "$SPARK_K8S_CMD" in
+ shift 1
+ SPARK_CLASSPATH=':/opt/spark/jars/*'
+ env
+ grep SPARK_JAVA_OPT_
+ + sed sort -t_ 's/[^=]*=\(.*\)/\1/g'-k4
-n
+ readarray -t SPARK_EXECUTOR_JAVA_OPTS
+ '[' -n '' ']'
+ '[' -n '' ']'
+ PYSPARK_ARGS=
+ '[' -n '' ']'
+ R_ARGS=
+ '[' -n '' ']'
+ '[' '' == 2 ']'
+ '[' '' == 3 ']'
+ case "$SPARK_K8S_CMD" in
+ CMD=("$SPARK_HOME/bin/spark-submit" --conf "spark.driver.bindAddress=$SPARK_DRIVER_BIND_ADDRESS" --deploy-mode client "$@")
+ exec /sbin/tini -s -- /opt/spark/bin/spark-submit --conf spark.driver.bindAddress=192.168.225.14 --deploy-mode client --properties-file /opt/spark/conf/spark.properties --class spark.jobs.app.streaming.Main spark-internal
20/04/20 09:27:20 WARN NativeCodeLoader: Unable to load native-hadoop library for your platform... using builtin-java classes where applicable
log4j:WARN No appenders could be found for logger (org.apache.spark.deploy.SparkSubmit$$anon$2).
log4j:WARN Please initialize the log4j system properly.
log4j:WARN See http://logging.apache.org/log4j/1.2/faq.html#noconfig for more info.
Pod события как показано с kubectl describe pod:
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal Scheduled 15m default-scheduler Successfully assigned spark/myapp-driver to xxxxxxxx.preprod.local
Warning FailedMount 15m kubelet, xxxxxxxx.preprod.local MountVolume.SetUp failed for volume "spark-conf-volume" : configmap "myapp-1587388343593-driver-conf-map" not found
Warning DNSConfigForming 15m (x4 over 15m) kubelet, xxxxxxxx.preprod.local Search Line limits were exceeded, some search paths have been omitted, the applied search line is: spark.svc.cluster.local svc.cluster.local cluster.local preprod.local
Normal Pulling 15m kubelet, xxxxxxxx.preprod.local Pulling image "xxxxxxxx:80/apache/myapp-test"
Normal Pulled 15m kubelet, xxxxxxxx.preprod.local Successfully pulled image "xxxxxxxx:80/apache/myapp-test"
Normal Created 15m kubelet, xxxxxxxx.preprod.local Created container spark-kubernetes-driver
Normal Started 15m kubelet, xxxxxxxx.preprod.local Started container spark-kubernetes-driver