Release 1.3.1

From the 1.3.1 milestone:

Enhancements:

• #45 Update commons-io:commons-io 2.7 -> 2.8.0
• #44 List files and directories by user (tool)
• #43 Show inode details (tool)
• #42 Tune fsimage file generation (fsimage generator)

Bugfixes:

• #41 Regression: Possible NPE when handling exception during section loading

1.3.0

From the 1.3.0 milestone:

This release contains major refactorings and breaking changes:

• Introduce builders for configurable loading/processing with different strategies (sequential/parallel etc)
• Split FSImageLoader into
  • FsImageLoader for loading file
  • FsImageData holding loaded data
  • FsVisitor and FsVisitorBuilder for sequential or parallel hierarchical access to FsImageData
• FSImageLoader.getFileSize -> FsUtil.getFileSize

Enhancements:

• #40 Support fsimages with compression
• #39 FSImage Generator : Support generating fsimages with compression
• #38 Cache root INode
• #37 Refactor inodes handling into abstraction
• #36 Update dependencies (eg Hadoop 3.2.0 -> 3.3.0)

1.2.0

From the 1.2.0 milestone:

Enhancements:

• #35 Optimize memory when loading FSImage
• #34 Update slf4j from 1.7.25 to latest 1.7.26
• #32 Warn when JVM heap seems insufficient for fsimage size (tool)
• #30 Support printing version for tool (tool)
• #31 Optimize Hadoop dependencies
• #29 Update dependency picocli to 3.9.6 (tool)
• #28 Support Hadoop 3 fsimage
• #19 Report and list top small files (tool)

1.1.5

From the 1.1.5 milestone:

Enhancements:

• #26 Support HDFS block storage policy with helper

Bugs:

• #27 Fix tool symlink reporting
• #25 Update guava dependency to avoid shade conflict

1.1.4

From the 1.1.4 milestone:

Enhancements:

• #18 Update dependencies
• #21 CLI option to toggle verbosity

Bugs:

• #24 Fix NPE on empty FSImage
• #23 FSImageLoader.visitParallel(...) passes wrong path to FsVisitor.onFile for root path
• #20 SizeBucket.computeBucketUpperBorders throws AIOOBE

Misc:

• #22 Migrate to assertj

1.1.3

From the 1.1.3 milestone:

Enhancements:

• #17 Increase buffered reader buffer size (8KiB ->64KiB)
• #16 Minor performance improvement : smarter section loading
• #15 Update com.google.guava:guava:21.0 to latest 24.0-jre
• #14 Minor performance tweak: Avoid Map.contains by optimistically getting and checking for null
• #13 Minor performance tweak to loading directories

1.1.2 Minor performance release

From the 1.1.2 milestone:

Enhancements:

• #12 : Improve visit performance by caching node ids for faster node lookups.
  Visits take about half the time now, for about 8MB more heap, per 1M inodes.
• #10 : Update depenencies
  hadoop 2.7.3 -> 2.7.5, commons-io 2.5 -> 2.6

1.1.1 Minor enhancements for consistency and bugfixes

From the 1.1.1 milestone:

Enhancements:

• #8 : Refactor IEC unit formatting to lib util package
• #7 : Support POSIX-like slashes in paths

Bugfixes:

• #9 : Tool throws AIOOBE bug

1.1 Performance release

Enhancements

• #1 Check if INode exists for given path
• #2 Support checking if inode/path has children
• #3 Update dependency picoli to 1.0.1
• #4 Tune performance : up to 3x faster by partial protobuf parsing

Bug fixes

• #5 SizeBucket.add throws IOOBE
• #6 Tool fails with MissingFormatArgumentException when autoextending buckets for very large files