Linux boot Kettle and Linux and Windows kettle to HDFs write data (3)

Xshell run into the graphical interface in xmanager

1 sh spoon. SH

Create a new job

1. write data into HDFs

1) kettle writes data to HDFs in Linux

Double-click hadoop copy Files

Run this job

View data:

1) kettle Write Data to HDFs in Windows

HDFs writes data to the power server in Windows

Log:

2016/07/28 16:21:14-version CHECKER-OK2016/07/28 16:21:57-Data integration tools-job designer-data integration tools-Job designer 2016/07/28 16:23:10-Data integration tools-jobs Designer-Save As ... 2016/07/28 16:23:10-Data integration tool-Job designer-Save file as ... 2016/07/28 16:23:17-Data integration tool-job designer-starting task ... 2016/07/28 16:23:17-ff-Start Task 2016/07/28 16:23:17-ff-Start Item [Hadoop copy files]2016/07/28 16:23:17-hadoop copy Files -Start ... 2016/07/28 16:23:17-hadoop Copy Files-processing rows, source files/directories: [File:///E:/auphi_designer_3.2.0/ez11.xls] ... Target file/directory: [Hdfs://172.16.1.20:8020/ke] ... Wildcard characters: [Null]2016/07/28 16:25:05-hadoop Copy files-error (version 3.2.0, build 1 from 2016-07-07 10.46.10 by Xnren): text System exception: Could not copy ' File:///E:/auphi_designer_3.2.0/ez11.xls ' to ' Hdfs://172.16.1.20:8020/ke/ez11.xls '. 2016/07/28 16:25:05-hadoop Copy files-error (version 3.2.0, build 1 from 2016-07-07 10.46.10 by Xnren): caused by:could not CLO Se the output stream for file "Hdfs://172.16.1.20:8020/ke/ez11.xls". 2016/07/28 16:25:05-hadoop Copy files-error (versi On 3.2.0, bUild 1 from 2016-07-07 10.46.10 by Xnren): caused by:connection timed out:no further INFORMATION2016/07/28 16:25:05-f F-Complete job Item [Hadoop Copy Files] (Result =[false]) 2016/07/28 16:25:05-ff-Task completed 2016/07/28 16:25:05-Data integration tool-job designer-the task has ended.

Cause of Error:

Because Windows is an external network, and the IP of all the machines in the hadoop cluster is intranet, although it can bind the external IPto the primary node, Windows can connect to the cluster master node but Write data to HDFs requires The intranet in the Hadoop cluster, so the data cannot be written in, Kettle migrated in the Linux Intranet environment, writing data successfully, as on 1).

Linux boot Kettle and Linux and Windows Kettle Write data to HDFs (3)