1. Sub-task HDFS-4028
    QJM: Merge newEpoch and prepareRecovery
  2. Sub-task HDFS-4025
    QJM: Sychronize past log segments to JNs that missed them
  3. Sub-task HDFS-4017
    Unclosed FileInputStream in GetJournalEditServlet
  4. Sub-task HDFS-4004
    TestJournalNode#testJournal fails because of test case execution order
  5. Sub-task HDFS-3956
    QJM: purge temporary files when no longer within retention period
  6. Sub-task HDFS-3955
    QJM: Make acceptRecovery() atomic
  7. Sub-task HDFS-3950
    QJM: misc TODO cleanup, improved log messages, etc
  8. Sub-task HDFS-3943
    QJM: remove currently unused "md5sum" field.
  9. Sub-task HDFS-3926
    QJM: Add user documentation for QJM
  10. Sub-task HDFS-3915
    QJM: Failover fails with auth error in secure cluster
  11. Sub-task HDFS-3914
    QJM: acceptRecovery should abort current segment
  12. Sub-task HDFS-3906
    QJM: quorum timeout on failover with large log segment
  13. Sub-task HDFS-3901
    QJM: send 'heartbeat' messages to JNs even when they are out-of-sync
  14. Sub-task HDFS-3900
    QJM: avoid validating log segments on log rolls
  15. Sub-task HDFS-3899
    QJM: Writer-side metrics
  16. Sub-task HDFS-3898
    QJM: enable TCP_NODELAY for IPC
  17. Bug HDFS-3897
    QJM: TestBlockToken fails after HDFS-3893
  18. Sub-task HDFS-3894
    QJM: testRecoverAfterDoubleFailures can be flaky due to IPC client caching
  19. Sub-task HDFS-3893
    QJM: Make QJM work with security enabled
  20. Sub-task HDFS-3891
    QJM: SBN fails if selectInputStreams throws RTE
  21. Sub-task HDFS-3885
    QJM: optimize log sync when JN is lagging behind
  22. Sub-task HDFS-3884
    QJM: Journal format() should reset cached values
  23. Sub-task HDFS-3877
    QJM: Provide defaults for dfs.journalnode.*address
  24. Sub-task HDFS-3870
    QJM: add metrics to JournalNode
  25. Sub-task HDFS-3869
    QJM: expose non-file journal manager details in web UI
  26. Sub-task HDFS-3863
    QJM: track last "committed" txid
  27. Sub-task HDFS-3845
    Fixes for edge cases in QJM recovery protocol
  28. Sub-task HDFS-3840
    JournalNodes log JournalNotFormattedException backtrace error before being formatted
  29. Sub-task HDFS-3839
    QJM: hadoop-daemon.sh should be updated to accept "journalnode"
  30. Improvement HDFS-3826
    QJM: Some trivial logging / exception text improvements
  31. Bug HDFS-3823
    QJM: TestQJMWithFaults fails occasionally because of missed setting of HTTP port
  32. Sub-task HDFS-3800
    QJM: improvements to QJM fault testing
  33. Sub-task HDFS-3799
    QJM: handle empty log segments during recovery
  34. Sub-task HDFS-3798
    Avoid throwing NPE when finalizeSegment() is called on invalid segment
  35. Sub-task HDFS-3797
    QJM: add segment txid as a parameter to journal() RPC
  36. Sub-task HDFS-3795
    QJM: validate journal dir at startup
  37. Sub-task HDFS-3793
    Implement genericized format() in QJM
  38. Bug HDFS-3773
    TestNNWithQJM fails after HDFS-3741
  39. Sub-task HDFS-3741
    QJM: exhaustive failure injection test for skipped RPCs
  40. Sub-task HDFS-3726
    QJM: if a logger misses an RPC, don't retry that logger until next segment
  41. Sub-task HDFS-3725
    Fix QJM startup when individual JNs have gaps
  42. Sub-task HDFS-3694
    QJM: Fix getEditLogManifest to fetch httpPort if necessary
  43. Sub-task HDFS-3693
    QJM: JNStorage should read its storage info even before a writer becomes active
  44. Sub-task HDFS-3692
    QJM: support purgeEditLogs() call to remotely purge logs
  45. New Feature HDFS-3077
    Quorum-based protocol for reading and writing edit logs
Refresh results