It was modified the algorithm to decide if a job should end successfully or not. I don't know if I'm removing the temporary directories before the last reduce task save the result in HDFS.
But I think that this error means that the reduce task couldn't save the data in HDFS. On 10 April 2012 18:56, Harsh J <[email protected]> wrote: > Pedro, > > Could we also know what was modified, since you claim it happens only > in the modified build? > > On Tue, Apr 10, 2012 at 9:15 PM, Pedro Costa <[email protected]> wrote: > > When I'm executing an MapReduce example on my modified Hadoop MapReduce, > > sometimes the reduce task give me this error, and the example doesn't > > finish: > > > > [code] > > 2012-04-10 11:32:38,110 INFO hdfs.DFSClient.closeInternal:3231 Could not > > complete file /user/output//part-r-00000 retrying.... > > [/code] > > > > This happens normally, when I'm executing an example after restarting JT > > and TT, without restarting the Namenode and the datanode. > > > > What this error means? > > > > Please notice that this happens only to my modified version and not the > > official. I modified the version 0.20.1. > > > > -- > > Best regards, > > > > -- > Harsh J > -- Best regards,
