Minor version release notes

1.4.6.2

Bug fixes

• Fixed Write-Ahead Log (WAL) corruption caused by Hadoop Distributed File System (HDFS) bugs.

• Fixed anomalies that occurred when writing data to HDFS. These anomalies could bring down an entire RegionServer with no recovery path.

• Fixed an issue where the system returned llegalStateException: No result & no exception instead of NotServingRegionException during region splitting.

• Fixed an issue where the read/write splitting configuration caused RegionServer launch failures.

• Fixed a bug where Phoenix Community Edition kept creating ZooKeeper connections until reaching the connection limit.

Improvements

• Optimized memory configurations on RegionServers to reduce the risk of out of memory (OOM) errors. Wide-scale scan memory consumption is significantly reduced.

• Optimized the kernel.

1.4.6.3

Bug fixes

• Fixed an issue where an index table created with IndexTool could not be found when the table name was specified in lowercase.

• Fixed an issue where a modified client time zone did not take effect. The default time zone is GTM.

• Fixed an NPE error that occurred when stale key-values were removed by replication after a column family was deleted.

Improvements

• Optimized default parameters for small-sized ApsaraDB for HBase clusters.

• Adjusted the maximum mutation size and bytes to prevent indexing failures on large cells.

• Replaced major compaction with raw scan to reduce the risk of index tables being disabled.

• Enforced index writes on failures to avoid disabling index tables. When index tables are disabled, the system falls back to scanning the entire primary table on queries.

• Disabled scan block cache during upsert-select and create index operations to reduce memory consumption and garbage collection overhead.

• Added port customization support to MiniHBaseCluster.

1.4.9.1

New features

• Added support for mixing cold data storage and hot data storage in the same ApsaraDB for HBase cluster.

• Added support for the Zstandard and LZO (Lempel-Ziv-Oberhumer) compression algorithms.

Bug fixes

• Fixed an issue where Multi-Version Concurrency Control (MVCC) stopped responding.

1.5.0

New features

• Added support for replication and synchronization between primary and secondary clusters.

Bug fixes

• Fixed a memory leak. This issue is reported to the Apache HBase community.

• Fixed an issue where the Reader thread was terminated on an out of memory (OOM) error without aborting the RegionServer. This issue is reported to the Apache HBase community.

1.5.1

Bug fixes

• Fixed compatibility issues caused by the Zstandard compression algorithm.

• Fixed performance issues in cold data storage.

• Fixed issues in Apache Phoenix secondary indexes.

• Fixed a memory leak. This issue is reported to the Apache HBase community.

• Fixed other bugs.

1.5.2

New features

• Added automatic locking when only a small amount of disk space remains.

Changes

• PREFIX_TREE encoding is not supported when creating tables.

1.5.3

New features

• Added a switch for dynamically enabling and disabling log subscription.

• Added support for monitoring active connections on the Thrift server.

Changes

• Disabled the Hbck command by default.

1.5.4

Improvements

• Limited the length of keys and values to prevent OOM errors caused by exceptional requests.

1.5.5

Bug fixes

• Fixed RegionServer failures caused by HLog Exporter.

1.5.5.1

Improvements

• Updated the logic for calculating the sizes of reference files.

2.0.2

New features

• Released ApsaraDB for HBase Enterprise Edition for commercial use.

• Upgraded ApsaraDB for HBase Basic Edition V1.1 to ApsaraDB for HBase V2.0.

Bug fixes

• Fixed a large number of bugs. These issues are reported to the Apache HBase community.

2.0.3

Bug fixes

• Fixed stability issues in AssignmentManager V2.

2.0.4

New features

• Added support for mixing cold data storage and hot data storage in the same ApsaraDB for HBase cluster.

Bug fixes

• Fixed stability issues.

2.0.5

Bug fixes

• Fixed the Apache Phoenix 5.x compatibility issue.

• Fixed performance issues in cold data storage.

2.0.6

New features

• Added support for the Zstandard compression algorithm.

Bug fixes

• Fixed stability issues. These issues are reported to the Apache HBase community.

2.0.7

New features

• Added automatic locking when only a small amount of disk space remains.

• Added support for log subscription and consumption from RegionServer.

Bug fixes

• Fixed an issue where the Phoenix WAL could not be cleared.

2.0.8

New features

• Added a switch for dynamically enabling and disabling log subscription.

Changes

• Disabled the Hbck command by default.

2.0.9

Bug fixes

• Fixed a memory leak in the DFS client.

• Fixed an issue where SplitTable blocked ModifyTable.

• Fixed WAL corruption that occurred after enabling ASYNC_WAL.

2.0.10

New features

• Added support for compaction interruption.

• Added the ability to limit the number of tables and regions supported by a RegionServer.

2.0.11

New features

• Added support for request throttling.

Bug fixes

• Fixed an issue where disable compaction did not take effect.

2.0.12

Bug fixes

• Fixed RegionServer failures caused by HLog Exporter.

Changes

• Disabled the Medium Object Storage (MOB) feature by default.

2.0.13

Bug fixes

• Fixed memory leaks caused by enabling mslab.