在Spark上使用S3（Frankfurt）

Question

Anyone is using s3 on Frankfurt using hadoop/spark 1.6.0?

I am trying to store the result of a job on s3, my dependencies are declared as follows:

"org.apache.spark" %% "spark-core" % "1.6.0" exclude("org.apache.hadoop", "hadoop-client"),
"org.apache.spark" %% "spark-sql" % "1.6.0",
"org.apache.hadoop" % "hadoop-client" % "2.7.2",
"org.apache.hadoop" % "hadoop-aws" % "2.7.2"

I have set the following configuration:

System.setProperty("com.amazonaws.services.s3.enableV4", "true")
sc.hadoopConfiguration.set("fs.s3a.endpoint", ""s3.eu-central-1.amazonaws.com")

When calling saveAsTextFile on my RDD it starts ok, saving everything on S3. However after some time when it is transferring from _temporary to the final output result it yields the error:

Exception in thread "main" com.amazonaws.services.s3.model.AmazonS3Exception: Status Code: 403, AWS Service: Amazon S3, AWS Request ID: XXXXXXXXXXXXXXXX, AWS Error Code: SignatureDoesNotMatch, AWS Error Message: The request signature we calculated does not match the signature you provided. Check your key and signing method., S3 Extended Request ID: XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX=
at com.amazonaws.http.AmazonHttpClient.handleErrorResponse(AmazonHttpClient.java:798)
at com.amazonaws.http.AmazonHttpClient.executeHelper(AmazonHttpClient.java:421)
at com.amazonaws.http.AmazonHttpClient.execute(AmazonHttpClient.java:232)
at com.amazonaws.services.s3.AmazonS3Client.invoke(AmazonS3Client.java:3528)
at com.amazonaws.services.s3.AmazonS3Client.copyObject(AmazonS3Client.java:1507)
at com.amazonaws.services.s3.transfer.internal.CopyCallable.copyInOneChunk(CopyCallable.java:143)
at com.amazonaws.services.s3.transfer.internal.CopyCallable.call(CopyCallable.java:131)
at com.amazonaws.services.s3.transfer.internal.CopyMonitor.copy(CopyMonitor.java:189)
at com.amazonaws.services.s3.transfer.internal.CopyMonitor.call(CopyMonitor.java:134)
at com.amazonaws.services.s3.transfer.internal.CopyMonitor.call(CopyMonitor.java:46)
at java.util.concurrent.FutureTask.run(FutureTask.java:266)
at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1142)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:617)
at java.lang.Thread.run(Thread.java:745)

If I use hadoop-client from spark package it not even start the transfer. The error occurs randomly, sometimes it works and sometimes don't.

Answer 1

Please try to set the values below:

System.setProperty("com.amazonaws.services.s3.enableV4", "true")
hadoopConf.set("fs.s3a.impl", "org.apache.hadoop.fs.s3a.S3AFileSystem")
hadoopConf.set("com.amazonaws.services.s3.enableV4", "true")
hadoopConf.set("fs.s3a.endpoint", "s3." + region + ".amazonaws.com")

please set the region where that bucket is located, in my case it was: eu-central-1

and add dependency into gradle or in some other way:

dependencies {
    compile 'org.apache.hadoop:hadoop-aws:2.7.2'
}

hope it will help.

Answer 2

In case you are using pyspark, the following worked for me

aws_profile = "your_profile"
aws_region = "eu-central-1"
s3_bucket = "your_bucket"

# see https://github.com/jupyter/docker-stacks/issues/127#issuecomment-214594895
os.environ['PYSPARK_SUBMIT_ARGS'] = "--packages=org.apache.hadoop:hadoop-aws:2.7.3 pyspark-shell"

# If this doesn't work you might have to delete your ~/.ivy2 directory to reset your package cache.
# (see https://github.com/databricks/spark-redshift/issues/244#issuecomment-239950148)
import pyspark
sc=pyspark.SparkContext()
# see https://github.com/databricks/spark-redshift/issues/298#issuecomment-271834485
sc.setSystemProperty("com.amazonaws.services.s3.enableV4", "true")

# see https://stackoverflow.com/questions/28844631/how-to-set-hadoop-configuration-values-from-pyspark
hadoop_conf=sc._jsc.hadoopConfiguration()
# see https://stackoverflow.com/questions/43454117/how-do-you-use-s3a-with-spark-2-1-0-on-aws-us-east-2
hadoop_conf.set("fs.s3a.impl", "org.apache.hadoop.fs.s3a.S3AFileSystem")
hadoop_conf.set("com.amazonaws.services.s3.enableV4", "true")
hadoop_conf.set("fs.s3a.access.key", access_id)
hadoop_conf.set("fs.s3a.secret.key", access_key)

# see https://docs.aws.amazon.com/general/latest/gr/rande.html#s3_region
hadoop_conf.set("fs.s3a.endpoint", "s3." + aws_region + ".amazonaws.com")

sql=pyspark.sql.SparkSession(sc)
path = s3_bucket + "your_file_on_s3"
dataS3=sql.read.parquet("s3a://" + path)

Answer 3

Inspired from the others answers, running the following directly in pyspark shell produced the desired output for me:

sc.setSystemProperty("com.amazonaws.services.s3.enableV4", "true") # fails without this
hc=sc._jsc.hadoopConfiguration()
hc.set("fs.s3a.impl", "org.apache.hadoop.fs.s3a.S3AFileSystem")
hc.set("com.amazonaws.services.s3.enableV4", "true")
hc.set("fs.s3a.endpoint", end_point)
hc.set("fs.s3a.access.key",access_key)
hc.set("fs.s3a.secret.key",secret_key)
data = sc.textFile("s3a://bucket/file")
data.take(3)

Choose your endpoint at: list of endpoints I was able to fetch data from Asia Pacific (Mumbai)(ap-south-1) which is a Version 4 only region.