hadoop 2.4.0 使用distcp有关问题解决

来源:互联网 发布:linux 禁ping 编辑:程序博客网 时间:2024/06/05 11:05

hadoop distcp hftp://nn.xxx.xx.com:50070/user/nlp/warehouse/t_m_user_key_action    /user/nlp/warehouse/dw1

出现

   Caused by: java.io.IOException: Check-sum mismatch between hftp://xxx:50070/foo/yyy.yy and hdfs://dst:8020/foo/xxx.xx



引用

— Distcp using MRv2 (YARN) from a CDH3 cluster to a CDH4 cluster may fail with CRC mismatch errors

Running distcp on a CDH4 YARN cluster with a CDH3 hftp source will fail if the CRC checksum type being used is the CDH4 default (CRC32C). This is because the default checksum type was changed in CDH4 from the CDH3 default of CRC32.

Bug: HADOOP-8060
Severity: Medium
Anticipated Resolution: To be fixed in an upcoming release
Workaround: You can work around this issue by changing the CRC checksum type on the CDH4 cluster to the CDH3 default, CRC32. To do this set dfs.checksum.type to CRC32 in hdfs-site.xml.


 在hdfs-site.xml文件里面添加:

   <property>
<name>dfs.checksum.type</name>
<value>CRC32</value>
</property>


注意执行命令的集群已经要有另一个集群的所有hosts文件。

0 0
原创粉丝点击