Uploaded image for project: 'Hadoop HDFS'
  1. Hadoop HDFS
  2. HDFS-3672

Expose disk-location information for blocks to enable better scheduling

    XMLWordPrintableJSON

Details

    • Improvement
    • Status: Closed
    • Major
    • Resolution: Fixed
    • 2.0.0-alpha
    • 2.0.2-alpha
    • None
    • None
    • Reviewed

    Description

      Currently, HDFS exposes on which datanodes a block resides, which allows clients to make scheduling decisions for locality and load balancing. Extending this to also expose on which disk on a datanode a block resides would enable even better scheduling, on a per-disk rather than coarse per-datanode basis.

      This API would likely look similar to Filesystem#getFileBlockLocations, but also involve a series of RPCs to the responsible datanodes to determine disk ids.

      Attachments

        1. hdfs-3672-9.patch
          68 kB
          Andrew Wang
        2. hdfs-3672-8.patch
          61 kB
          Andrew Wang
        3. hdfs-3672-7.patch
          61 kB
          Andrew Wang
        4. hdfs-3672-6.patch
          60 kB
          Andrew Wang
        5. hdfs-3672-5.patch
          52 kB
          Andrew Wang
        6. hdfs-3672-4.patch
          52 kB
          Andrew Wang
        7. hdfs-3672-3.patch
          49 kB
          Andrew Wang
        8. hdfs-3672-2.patch
          48 kB
          Andrew Wang
        9. hdfs-3672-12.patch
          69 kB
          Andrew Wang
        10. hdfs-3672-11.patch
          68 kB
          Andrew Wang
        11. hdfs-3672-10.patch
          68 kB
          Andrew Wang
        12. hdfs-3672-1.patch
          37 kB
          Andrew Wang
        13. design-doc-v2.pdf
          73 kB
          Andrew Wang
        14. design-doc-v1.pdf
          73 kB
          Andrew Wang

        Issue Links

          is depended upon by

          Brainstorming - A place to record back and forth on notions not yet formed enough to make a 'New Feature' or 'Task' HBASE-6572 Tiered HFile storage

          • Major - Major loss of function.
          • Closed
          is related to

          Bug - A problem which impairs or prevents the functions of the product. HDFS-3969 Small bug fixes and improvements for disk locations API

          • Major - Major loss of function.
          • Closed

          Bug - A problem which impairs or prevents the functions of the product. MAPREDUCE-4577 HDFS-3672 broke TestCombineFileInputFormat.testMissingBlocks() test

          • Minor - Minor loss of function, or other problem where easy workaround is present.
          • Closed

          New Feature - A new feature of the product, which has yet to be developed. HDFS-2832 Enable support for heterogeneous storages in HDFS - DN as a collection of storages

          • Major - Major loss of function.
          • Closed

          Improvement - An improvement or enhancement to an existing feature or task. HDFS-8895 Remove deprecated BlockStorageLocation APIs

          • Major - Major loss of function.
          • Resolved

          Brainstorming - A place to record back and forth on notions not yet formed enough to make a 'New Feature' or 'Task' HBASE-6572 Tiered HFile storage

          • Major - Major loss of function.
          • Closed
          is superceded by

          Improvement - An improvement or enhancement to an existing feature or task. HDFS-8887 Expose storage type and storage ID in BlockLocation

          • Major - Major loss of function.
          • Resolved

          Activity

            People

              andrew.wang Andrew Wang
              andrew.wang Andrew Wang
              Votes:
              0 Vote for this issue
              Watchers:
              30 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: