Hi there,
Can you check if HADOOP_CONF_DIR is being set on the executors to
/opt/spark/conf? One should set an executor environment variable for that.
A kubectl describe pod output for the executors would be helpful here.
-Matt Cheah
From: Oscar Bonilla <oscar.boni...@gmail.com>
Date: Friday, October 19, 2018 at 1:03 AM
To: "user@spark.apache.org" <user@spark.apache.org>
Subject: [Spark for kubernetes] Azure Blob Storage credentials issue
Hello,
I'm having the following issue while trying to run Spark for kubernetes
[spark.apache.org]:
2018-10-18 08:48:54 INFO DAGScheduler:54 - Job 0 failed: reduce at
SparkPi.scala:38, took 1.743177 s
Exception in thread "main" org.apache.spark.SparkException: Job aborted due to
stage failure: Task 0 in stage 0.0 failed 4 times, most recent failure: Lost
task 0.3 in stage 0.0 (TID 6, 10.244.1.11, executor 2):
org.apache.hadoop.fs.azure.AzureException:
org.apache.hadoop.fs.azure.AzureException: No credentials found for account
datasets83d858296fd0c49b.blob.core.windows.net
[datasets83d858296fd0c49b.blob.core.windows.net] in the configuration, and its
container datasets is not accessible using anonymous credentials. Please check
if the container exists first. If it is not publicly available, you have to
provide account credentials.
at
org.apache.hadoop.fs.azure.AzureNativeFileSystemStore.createAzureStorageSession(AzureNativeFileSystemStore.java:1086)
at
org.apache.hadoop.fs.azure.AzureNativeFileSystemStore.initialize(AzureNativeFileSystemStore.java:538)
at
org.apache.hadoop.fs.azure.NativeAzureFileSystem.initialize(NativeAzureFileSystem.java:1366)
at org.apache.hadoop.fs.FileSystem.createFileSystem(FileSystem.java:3242)
at org.apache.hadoop.fs.FileSystem.access$200(FileSystem.java:121)
at org.apache.hadoop.fs.FileSystem$Cache.getInternal(FileSystem.java:3291)
at org.apache.hadoop.fs.FileSystem$Cache.get(FileSystem.java:3259)
at org.apache.hadoop.fs.FileSystem.get(FileSystem.java:470)
at org.apache.spark.util.Utils$.getHadoopFileSystem(Utils.scala:1897)
at org.apache.spark.util.Utils$.doFetchFile(Utils.scala:694)
at org.apache.spark.util.Utils$.fetchFile(Utils.scala:476)
at
org.apache.spark.executor.Executor$$anonfun$org$apache$spark$executor$Executor$$updateDependencies$5.apply(Executor.scala:755)
at
org.apache.spark.executor.Executor$$anonfun$org$apache$spark$executor$Executor$$updateDependencies$5.apply(Executor.scala:747)
at
scala.collection.TraversableLike$WithFilter$$anonfun$foreach$1.apply(TraversableLike.scala:733)
at
scala.collection.mutable.HashMap$$anonfun$foreach$1.apply(HashMap.scala:99)
at
scala.collection.mutable.HashMap$$anonfun$foreach$1.apply(HashMap.scala:99)
at
scala.collection.mutable.HashTable$class.foreachEntry(HashTable.scala:230)
at scala.collection.mutable.HashMap.foreachEntry(HashMap.scala:40)
at scala.collection.mutable.HashMap.foreach(HashMap.scala:99)
at
scala.collection.TraversableLike$WithFilter.foreach(TraversableLike.scala:732)
at org.apache.spark.executor.Executor.org
[org.apache.spark.executor.executor.org]$apache$spark$executor$Executor$$updateDependencies(Executor.scala:747)
at org.apache.spark.executor.Executor$TaskRunner.run(Executor.scala:312)
at
java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1149)
at
java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:624)
at java.lang.Thread.run(Thread.java:748)
Caused by: org.apache.hadoop.fs.azure.AzureException: No credentials found for
account datasets83d858296fd0c49b.blob.core.windows.net
[datasets83d858296fd0c49b.blob.core.windows.net] in the configuration, and its
container datasets is not accessible using anonymous credentials. Please check
if the container exists first. If it is not publicly available, you have to
provide account credentials.
at
org.apache.hadoop.fs.azure.AzureNativeFileSystemStore.connectUsingAnonymousCredentials(AzureNativeFileSystemStore.java:863)
at
org.apache.hadoop.fs.azure.AzureNativeFileSystemStore.createAzureStorageSession(AzureNativeFileSystemStore.java:1081)
... 24 more
The command I use to launch the job is:
/opt/spark/bin/spark-submit
--master k8s://<my-k8s-master>
--deploy-mode cluster
--name spark-pi
--class org.apache.spark.examples.SparkPi
--conf spark.executor.instances=5
--conf spark.kubernetes.container.image=<my-image-built-with-wasb>
--conf spark.kubernetes.namespace=<my-namespace>
--conf spark.kubernetes.authenticate.driver.serviceAccountName=spark
--conf spark.kubernetes.driver.secrets.spark=/opt/spark/conf
--conf spark.kubernetes.executor.secrets.spark=/opt/spark/conf
wasb://<my-container-name>@<my-account-name>.blob.core.windows.net/spark-examples_2.11-2.3.2.jar
[blob.core.windows.net] 10000
I have a k8s secret named spark with the following content:
apiVersion: v1
kind: Secret
metadata:
name: spark
labels:
app: spark
stack: service
type: Opaque
data:
core-site.xml: |-
{% filter b64encode %}
<configuration>
<property>
<name>fs.azure.account.key.<my-account-name>.blob.core.windows.net
[blob.core.windows.net]</name>
<value><my-account-key></value>
</property>
<property>
<name>fs.AbstractFileSystem.wasb.Impl</name>
<value>org.apache.hadoop.fs.azure.Wasb</value>
</property>
</configuration>
{% endfilter %}
The driver pod manages to download the jar dependencies as stored in a
container in Azure Blob Storage. As can be seen in this log snippet:
2018-10-18 08:48:16 INFO Utils:54 - Fetching
wasb://<my-container-name>@<my-account-name>.blob.core.windows.net/spark-examples_2.11-2.3.2.jar
[blob.core.windows.net] to
/var/spark-data/spark-jars/fetchFileTemp8575879929413871510.tmp
2018-10-18 08:48:16 INFO SparkPodInitContainer:54 - Finished downloading
application dependencies.
How can I get the executor pods to get the credentials as stored in the
core-site.xml file that's mounted from the k8s secret? What am I missing?
Thank you,
Oscar Bonilla.
smime.p7s
Description: S/MIME cryptographic signature
