Uploaded image for project: 'Hadoop HDFS'
  1. Hadoop HDFS
  2. HDFS-2139 Fast copy for HDFS.
  3. HDFS-17592

FastCopy support data copy in different nameservices without federation

    XMLWordPrintableJSON

Details

    • Sub-task
    • Status: Open
    • Major
    • Resolution: Unresolved
    • None
    • None
    • None
    • None

    Description

      FastCopy is  a faster data copy tools.  In federation cluster  or a single cluster , FastCopy copy blocks via hardlink.  This is more much faster than original copy.

      FastCopy can support data copy via transfer in different nameservices without federation. In theory, it could reduces one IO transfer and almost reduce halt time.

       

      Test Data:

      blocksize 128M

      1TB ECfiles + 1TB 3 replicated files

       

      distcp map=20 DIstcp via FastCopy(HardLink) DistCp via FastCopy(Transfer) Distcp(original)
       Time Spent 5m6.687s 22m44.094s 38m17.024s

       

      Attachments

        1. FastCopy via Transfer.jpg
          52 kB
          liuguanghua

        Activity

          People

            liuguanghua liuguanghua
            liuguanghua liuguanghua
            Votes:
            0 Vote for this issue
            Watchers:
            1 Start watching this issue

            Dates

              Created:
              Updated: