Develop a good project in eclipse , the program has two job, so write a scheduling class JobSchedule to control the order of execution of these two job , as well as dealing with some inputpath and stuff .
The whole process successful implementation in eclipse , no problem , but when the program packaged into jar, get linux, bin / hadoop below to perform when the log output on and executed in eclipse is not the same In addition to the program execution process is not moving. . Not sure how else
execution in linux command bin / hadoop hadoop jar MyJob.jar
because the input is the output path of the directory structure in the program designated , so there is no path with the input and output parameters.
there who encounter such a situation, or know the reason, brother Kouxie a ~ ~
------ Solution ----------------- ---------------------------
only own analysis of the error log of
------ Solution --- -----------------------------------------
not met , no too well
packaged runnable jar file when the election try it .
------ For reference only -------------------------------------- -
people do ~
------ For reference only ---------------------------------------
ah . Problem solved ~ ~ NND, is the need to put all the machine's IP address and host name configuration to / etc / hosts file to go, another master machine's / etc / hosts file first line 127.0.0.1 .. .. this should be commented out ~ ~
Thank you good people for help or friends ~ ~
没有评论:
发表评论