Uploaded image for project: 'Hadoop HDFS'
  1. Hadoop HDFS
  2. HDFS-142

In 0.20, move blocks being written into a blocksBeingWritten directory

    XMLWordPrintableJSON

Details

    • Bug
    • Status: Closed
    • Blocker
    • Resolution: Fixed
    • 0.20-append
    • 0.20-append, 0.20.205.0
    • None
    • None
    • hbase

    Description

      Before 0.18, when Datanode restarts, it deletes files under data-dir/tmp directory since these files are not valid anymore. But in 0.18 it moves these files to normal directory incorrectly making them valid blocks. One of the following would work :

      • remove the tmp files during upgrade, or
      • if the files under /tmp are in pre-18 format (i.e. no generation), delete them.

      Currently effect of this bug is that, these files end up failing block verification and eventually get deleted. But cause incorrect over-replication at the namenode before that.

      Also it looks like our policy regd treating files under tmp needs to be defined better. Right now there are probably one or two more bugs with it. Dhruba, please file them if you rememeber.

      Attachments

        1. deleteTmp.patch
          6 kB
          Dhruba Borthakur
        2. deleteTmp2.patch
          13 kB
          Dhruba Borthakur
        3. deleteTmp_0.18.patch
          2 kB
          Dhruba Borthakur
        4. handleTmp1.patch
          9 kB
          Dhruba Borthakur
        5. appendQuestions.txt
          8 kB
          Dhruba Borthakur
        6. deleteTmp5_20.txt
          35 kB
          Dhruba Borthakur
        7. deleteTmp5_20.txt
          32 kB
          Dhruba Borthakur
        8. HDFS-142_20.patch
          54 kB
          Nicolas Spiegelberg
        9. hdfs-142-minidfs-fix-from-409.txt
          2 kB
          Todd Lipcon
        10. HDFS-142-multiple-blocks-datanode-exception.patch
          3 kB
          Karthik Ranganathan
        11. testfileappend4-deaddn.txt
          2 kB
          Todd Lipcon
        12. HDFS-142-deaddn-fix.patch
          3 kB
          Nicolas Spiegelberg
        13. HDFS-142-finalize-fix.txt
          7 kB
          sam rash
        14. hdfs-142-commitBlockSynchronization-unknown-datanode.txt
          2 kB
          Todd Lipcon
        15. hdfs-142-testcases.txt
          8 kB
          Todd Lipcon
        16. hdfs-142-recovery-reassignment-and-bbw-cleanup.txt
          19 kB
          Todd Lipcon
        17. hdfs-142-testleaserecovery-fix.txt
          5 kB
          Todd Lipcon
        18. validateBlockMetaData-synchronized.txt
          0.9 kB
          Todd Lipcon
        19. recentInvalidateSets-assertion-fix.txt
          1 kB
          Todd Lipcon
        20. appendFile-recheck-lease.txt
          6 kB
          Todd Lipcon
        21. dont-recover-rwr-when-rbw-available.txt
          24 kB
          Todd Lipcon
        22. recover-rbw-v2.txt
          27 kB
          Todd Lipcon
        23. HDFS-142_20-append2.patch
          84 kB
          Nicolas Spiegelberg
        24. HDFS-142.20-security.1.patch
          88 kB
          Jitendra Nath Pandey
        25. HDFS-142.20-security.2.patch
          89 kB
          Jitendra Nath Pandey

        Issue Links

          depends upon

          Bug - A problem which impairs or prevents the functions of the product. HDFS-101 DFS write pipeline : DFSClient sometimes does not detect second datanode failure

          • Blocker - Blocks development and/or testing work, production could not run
          • Closed

          Bug - A problem which impairs or prevents the functions of the product. HDFS-793 DataNode should first receive the whole packet ack message before it constructs and sends its own ack message for the packet

          • Blocker - Blocks development and/or testing work, production could not run
          • Closed

          Bug - A problem which impairs or prevents the functions of the product. HDFS-988 saveNamespace race can corrupt the edits log

          • Blocker - Blocks development and/or testing work, production could not run
          • Closed

          Bug - A problem which impairs or prevents the functions of the product. HDFS-606 ConcurrentModificationException in invalidateCorruptReplicas()

          • Major - Major loss of function.
          • Closed

          Improvement - An improvement or enhancement to an existing feature or task. HDFS-826 Allow a mechanism for an application to detect that datanode(s) have died in the write pipeline

          • Major - Major loss of function.
          • Closed
          incorporates

          Bug - A problem which impairs or prevents the functions of the product. HADOOP-4997 workaround for tmp file handling on DataNodes in 0.18 (HADOOP-4663)

          • Blocker - Blocks development and/or testing work, production could not run
          • Closed
          relates to

          Bug - A problem which impairs or prevents the functions of the product. HDFS-57 A Datanode's datadir could have lots of blocks in the top-level directory

          • Major - Major loss of function.
          • Resolved

          Bug - A problem which impairs or prevents the functions of the product. HADOOP-4702 Failed block replication leaves an incomplete block in receiver's tmp data directory

          • Blocker - Blocks development and/or testing work, production could not run
          • Closed

          Bug - A problem which impairs or prevents the functions of the product. HADOOP-4810 Data lost at cluster startup time

          • Blocker - Blocks development and/or testing work, production could not run
          • Closed

          Bug - A problem which impairs or prevents the functions of the product. HDFS-29 In Datanode, update block may fail due to length inconsistency

          • Major - Major loss of function.
          • Closed

          Activity

            People

              dhruba Dhruba Borthakur
              rangadi Raghu Angadi
              Votes:
              0 Vote for this issue
              Watchers:
              8 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: