python - 从 s3 ://to local(hadoop) file system 复制文件时出错

Question

我正在尝试使用 python 将文件从 s3 复制到 hadoop 文件系统。我收到以下错误:

cp: `foo/ds=2015-02-13/ip-d1b-request-2015-02-13_10-00_10-09.txt.gz': No such file or directory

我最近正在迁移最新的 hadoop 版本 (2.4.0)。在版本(0.20)中工作正常。为什么我在 2.4.0 版本中会出现此错误？

在 Hadoop 版本 0.20 中

hadoop@ip-10-76-38-167:~$ /home/hadoop/bin/hadoop fs -cp s3://test.com/foo/ds=2015-02-13/ip-d1b-request-2015-02-13_10-00_10-09.txt.gz /foo/ds=2015-02-13/ip-d1b-request-2015-02-13_10-00_10-09.txt.gz

15/02/13 11:21:45 INFO s3native.NativeS3FileSystem: Opening 's3://test.com/foo/ds=2015-02-13/ip-d1b-request-2015-02-13_10-00_10-09.txt.gz' for reading

在 Hadoop 版本 2.4.0 中

[hadoop@ip-10-169-19-123 ~]$ /home/hadoop/bin/hadoop fs -cp s3://test.com/foo/ds=2015-02-13/ip-d1b-request-2015-02-13_10-00_10-09.txt.gz /foo/ds=2015-02-13/ip-d1b-request-2015-02-13_10-00_10-09.txt.gz

15/02/13 11:21:37 INFO guice.EmrFSBaseModule: Consistency disabled, using com.amazon.ws.emr.hadoop.fs.s3n.S3NativeFileSystem as FileSystem implementation.

15/02/13 11:21:38 INFO fs.EmrFileSystem: Using com.amazon.ws.emr.hadoop.fs.s3n.S3NativeFileSystem as filesystem implementation

cp: `foo/ds=2015-02-13/ip-d1b-request-2015-02-13_10-00_10-09.txt.gz': No such file or directory

Answer 1

你需要这样尝试。添加“

/home/hadoop/bin/hadoop fs -cp "s3://test.com/foo/ds=2015-02-13/ip-d1b-request-2015-02-13_10-00_10-09.txt.gz" "/foo/ds=2015-02-13/ip-d1b-request-2015-02-13_10-00_10-09.txt.gz"