Details
-
Improvement
-
Status: Resolved
-
Minor
-
Resolution: Not A Problem
-
None
-
None
-
None
-
from 35 node cluster to 10 node cluster
Description
When starting a distcp the setup phase often takes a very long time. For example during the distcp I just ran the setup phase took 15 minutes and the actual copy 3 minutes. Could this be improved? Or at least a progress bar added so the user doesn't think it stalled.
I also often see exceptions like this in the setup, but the distcp finishes eventually.
java.io.EOFException
at java.io.DataInputStream.readShort(DataInputStream.java:298)
at org.apache.hadoop.dfs.DFSClient$DFSOutputStream.endBlock(DFSClient.java:1672)
at org.apache.hadoop.dfs.DFSClient$DFSOutputStream.close(DFSClient.java:1744)
at org.apache.hadoop.fs.FSDataOutputStream$PositionCache.close(FSDataOutputStream.java:49)
at org.apache.hadoop.fs.FSDataOutputStream.close(FSDataOutputStream.java:64)
at org.apache.hadoop.io.SequenceFile$Writer.close(SequenceFile.java:774)
at org.apache.hadoop.util.CopyFiles$FSCopyFilesMapper.setup(CopyFiles.java:351)
at org.apache.hadoop.util.CopyFiles.copy(CopyFiles.java:773)
at org.apache.hadoop.util.CopyFiles.run(CopyFiles.java:854)
at org.apache.hadoop.util.ToolBase.doMain(ToolBase.java:187)
at org.apache.hadoop.util.CopyFiles.main(CopyFiles.java:864)