403 与 hadoop-aws 和 s3proxy

Question

我正在尝试将 s3proxy 与 Hadoop 3.1.2 一起使用，但出现 403 错误，而在 hadoop 2.7 上它似乎可以工作。我这边是倒退还是配置错误？

s3proxy 配置

s3proxy.endpoint=http://127.0.0.1:4242
s3proxy.authorization=aws-v2
s3proxy.identity=local-identity
s3proxy.credential=local-credential
jclouds.provider=filesystem
jclouds.filesystem.basedir=/tmp/s3proxy

简单的 Scala 代码

sc.hadoopConfiguration.set("fs.s3a.impl", "org.apache.hadoop.fs.s3a.S3AFileSystem")
sc.hadoopConfiguration.set("fs.s3a.aws.credentials.provider", "org.apache.hadoop.fs.s3a.SimpleAWSCredentialsProvider")
sc.hadoopConfiguration.set("fs.s3a.endpoint", "127.0.0.1:4242")
sc.hadoopConfiguration.set("fs.s3a.connection.ssl.enabled", "false")
sc.hadoopConfiguration.set("fs.s3a.access.key", "local-identity")
sc.hadoopConfiguration.set("fs.s3a.secret.key", "local-credential")
val rdd = sc.textFile("s3a://wiki/test.json")
rdd.collect().foreach(println)

和tree /tmp/s3proxy

/tmp/s3proxy/
└── wiki
    └── test.json

使用 hadoop2.7，我得到了正确的输出。

cat demo_wiki/test01.scala | ./spark-2.4.4-bin-hadoop2.7/bin/spark-shell --jars hadoop-aws-2.7.3.jar,aws-java-sdk-1.7.4.jar

但是对于 hadoop3.1.2，我遇到了 403 错误。

cat demo_wiki/test01.scala | ./spark-2.4.4-bin-hadoop3.1/bin/spark-shell --jars hadoop-aws-3.1.2.jar,aws-java-sdk-bundle-1.11.271.jar

输出：

2019-09-23 15:23:32 WARN  MetricsConfig:134 - Cannot locate configuration: tried hadoop-metrics2-s3a-file-system.properties,hadoop-metrics2.properties
java.nio.file.AccessDeniedException: s3a://wiki/test.json: getFileStatus on s3a://wiki/test.json: com.amazonaws.services.s3.model.AmazonS3Exception: Forbidden (Service: Amazon S3; Status Code: 403; Error Code: 403 Forbidden; Request ID: 4442587FB7D0A2F9; S3 Extended Request ID: null), S3 Extended Request ID: null:403 Forbidden

有什么想法吗？

Answer 1

查看 Dockerfile 找到答案 (https://github.com/gaul/s3proxy/blob/master/Dockerfile#L24)

正确的 s3proxy 配置是

s3proxy.endpoint=http://127.0.0.1:4242
s3proxy.authorization=aws-v2-or-v4
s3proxy.identity=local-identity
s3proxy.credential=local-credential
jclouds.provider=filesystem
jclouds.filesystem.basedir=/tmp/s3proxy

最新（当时）版本的 spark-2.4.4 + hadoop-3.2.0 + aws-sdk-1.11.636 的命令行运行良好

cat test.scala | ./spark-2.4.4-bin-hadoop3.2/bin/spark-shell --jars hadoop-aws-3.2.0.jar,aws-java-sdk-bundle-1.11.636.jar

403 与 hadoop-aws 和 s3proxy

403 with hadoop-aws and s3proxy

hadoop

amazon-s3

aws-sdk

s3proxy