The path indicates that its looking at hdfs. If you are loading data which is on your local file system, you should do something like LOAD DATA LOCAL INPATH '/path/in/local/fs'.
On 7/8/09 4:25 PM, "RSD" <[email protected]> wrote: > I've gotten new equipment to do an upgrade, but I need to keep my Hadoop > cluster > pushing data. :-) > > I am getting the following: > > Job Submission failed with exception 'Input path doesnt exist : > hdfs://XXXXXXXXXXXXXXX:9000/tmp/hive-username/183603784.10002' > FAILED: Execution Error, return code 1 from > org.apache.hadoop.hive.ql.exec.ExecDriver > > My hive.log: > 009-07-08 16:08:49,559 INFO exec.ExecDriver > (ExecDriver.java:execute(141)) - Adding input file > /tmp/hive-username/183603784.10002 > 2009-07-08 16:08:49,705 ERROR exec.ExecDriver > (SessionState.java:printError(242)) - Job Submission failed with > exception 'Input path doesnt exist : > hdfs://XXXXXXXXXXXXXXX:9000/tmp/hive-username/183603784.10002' > org.apache.hadoop.mapred.InvalidInputException: Input path doesnt > exist : hdfs://XXXXXXXXXXXXXXX:9000/tmp/hive-username/183603784.10002 > > > What does this error indicate? The files are in Hadoop that I want to > work on; I've checked permissions, as my first thought was that > I could not write to /tmp on the local machine, but that is not the > case. Any one have a clue?
