Hadoop HDFS
  1. Hadoop HDFS
  2. HDFS-1300

Decommissioning nodes does not increase replication priority

    Details

    • Type: Bug Bug
    • Status: Patch Available
    • Priority: Major Major
    • Resolution: Unresolved
    • Affects Version/s: 0.20.1, 0.20.2, 0.20.3, 0.20-append, 0.21.0, 0.22.0
    • Fix Version/s: None
    • Component/s: None
    • Labels:
      None

      Description

      Currently when you decommission a node each block is only inserted into neededReplications if it is not there yet. This causes a problem of a block sitting in a low priority queue when all replicas sit on the nodes being decommissioned.
      The common usecase for decommissioning nodes for us is proactively exclude them before they went bad, so it would be great to get the blocks at risk onto the live datanodes as quickly as possible.

      1. HDFS-1300.patch
        7 kB
        Dmytro Molkov
      2. HDFS-1300.3.patch
        7 kB
        Dmytro Molkov
      3. HDFS-1300.2.patch
        1 kB
        Dmytro Molkov

        Activity

        Hide
        Hadoop QA added a comment -

        -1 overall. Here are the results of testing the latest attachment
        http://issues.apache.org/jira/secure/attachment/12459401/HDFS-1300.3.patch
        against trunk revision 1072023.

        +1 @author. The patch does not contain any @author tags.

        +1 tests included. The patch appears to include 5 new or modified tests.

        +1 javadoc. The javadoc tool did not generate any warning messages.

        +1 javac. The applied patch does not increase the total number of javac compiler warnings.

        +1 findbugs. The patch does not introduce any new Findbugs (version 1.3.9) warnings.

        +1 release audit. The applied patch does not increase the total number of release audit warnings.

        -1 core tests. The patch failed these core unit tests:
        org.apache.hadoop.cli.TestHDFSCLI
        org.apache.hadoop.hdfs.server.datanode.TestBlockRecovery
        org.apache.hadoop.hdfs.TestFileConcurrentReader

        -1 contrib tests. The patch failed contrib unit tests.

        +1 system test framework. The patch passed system test framework compile.

        Test results: https://hudson.apache.org/hudson/job/PreCommit-HDFS-Build/191//testReport/
        Findbugs warnings: https://hudson.apache.org/hudson/job/PreCommit-HDFS-Build/191//artifact/trunk/build/test/findbugs/newPatchFindbugsWarnings.html
        Console output: https://hudson.apache.org/hudson/job/PreCommit-HDFS-Build/191//console

        This message is automatically generated.

        Show
        Hadoop QA added a comment - -1 overall. Here are the results of testing the latest attachment http://issues.apache.org/jira/secure/attachment/12459401/HDFS-1300.3.patch against trunk revision 1072023. +1 @author. The patch does not contain any @author tags. +1 tests included. The patch appears to include 5 new or modified tests. +1 javadoc. The javadoc tool did not generate any warning messages. +1 javac. The applied patch does not increase the total number of javac compiler warnings. +1 findbugs. The patch does not introduce any new Findbugs (version 1.3.9) warnings. +1 release audit. The applied patch does not increase the total number of release audit warnings. -1 core tests. The patch failed these core unit tests: org.apache.hadoop.cli.TestHDFSCLI org.apache.hadoop.hdfs.server.datanode.TestBlockRecovery org.apache.hadoop.hdfs.TestFileConcurrentReader -1 contrib tests. The patch failed contrib unit tests. +1 system test framework. The patch passed system test framework compile. Test results: https://hudson.apache.org/hudson/job/PreCommit-HDFS-Build/191//testReport/ Findbugs warnings: https://hudson.apache.org/hudson/job/PreCommit-HDFS-Build/191//artifact/trunk/build/test/findbugs/newPatchFindbugsWarnings.html Console output: https://hudson.apache.org/hudson/job/PreCommit-HDFS-Build/191//console This message is automatically generated.
        Hide
        Dmytro Molkov added a comment -

        submitting for hudson

        Show
        Dmytro Molkov added a comment - submitting for hudson
        Hide
        Dmytro Molkov added a comment -

        The patch with the test in it. Also fixed the test to work with the latest change in trunk.

        Show
        Dmytro Molkov added a comment - The patch with the test in it. Also fixed the test to work with the latest change in trunk.
        Hide
        Jakob Homan added a comment -

        Dmytro- Looks like the updated patch dropped the unit test. Can you upload a new one with it re-attached? Thanks.

        Show
        Jakob Homan added a comment - Dmytro- Looks like the updated patch dropped the unit test. Can you upload a new one with it re-attached? Thanks.
        Hide
        Dmytro Molkov added a comment -

        Submitting to Hudson again.

        Show
        Dmytro Molkov added a comment - Submitting to Hudson again.
        Hide
        Dmytro Molkov added a comment -

        Patch agains updated trunk

        Show
        Dmytro Molkov added a comment - Patch agains updated trunk
        Hide
        Dmytro Molkov added a comment -

        Submitting to Hudson.
        Can someone look at the code to see that it is OK?

        Show
        Dmytro Molkov added a comment - Submitting to Hudson. Can someone look at the code to see that it is OK?
        Hide
        Dmytro Molkov added a comment -

        Please have a look at the patch.
        Instead of inserting only when it doesn't exist we should rather follow the same way we handle dead datanodes. Update the block, so it will move up in priority queues when more nodes get decommissioned.

        Show
        Dmytro Molkov added a comment - Please have a look at the patch. Instead of inserting only when it doesn't exist we should rather follow the same way we handle dead datanodes. Update the block, so it will move up in priority queues when more nodes get decommissioned.

          People

          • Assignee:
            Dmytro Molkov
            Reporter:
            Dmytro Molkov
          • Votes:
            0 Vote for this issue
            Watchers:
            4 Start watching this issue

            Dates

            • Created:
              Updated:

              Development