Release Notes - Hadoop HBase - Version 0.20.3 - HTML format

Sub-task

• [HBASE-1982] - [EC2] handle potentially large and uneven instance startup times
• [HBASE-1993] - Set up convenient ssh between instances
• [HBASE-2009] - [EC2] support mapreduce subsystem
• [HBASE-2012] - [EC2] LZO support
• [HBASE-2032] - [EC2] Add support for installation of user packages when instances start up
• [HBASE-2064] - Cannot disable a table if at the same the Master is moving its regions around
• [HBASE-2065] - Cannot disable a table if any of its region is opening at the same time
• [HBASE-2067] - Disabling a table can return too soon if one region is pending open
• [HBASE-2080] - [EC2] Support multivolume local instance storage
• [HBASE-2083] - [EC2] HDFS DataNode no longer required on master
• [HBASE-2084] - [EC2] JAVA_HOME handling broken
• [HBASE-2100] - [EC2] Adjust fs.file-max
• [HBASE-2103] - [EC2] pull version from build
• [HBASE-2131] - [EC2] Mount data volumes as xfs, noatime
• [HBASE-2132] - [EC2] Rebuild AMIs with jdk6u18

Bug

• [HBASE-1636] - disable and drop of table is flakey still
• [HBASE-1813] - Remove compaction governor
• [HBASE-1960] - Master should wait for DFS to come up when creating hbase.version
• [HBASE-1962] - Bulk loading script makes regions incorrectly (loadtable.rb)
• [HBASE-1979] - MurmurHash does not yield the same results as the reference C++ implementation when size % 4 >= 2
• [HBASE-1994] - Master will lose hlog entries while splitting if region has empty oldlogfile.log
• [HBASE-1997] - zk tick time bounds maximum zk session time
• [HBASE-1999] - When HTable goes away, close zk session in shutdown hook or something...
• [HBASE-2003] - [shell] deleteall ignores column if specified
• [HBASE-2018] - Updates to .META. blocked under high MemStore load
• [HBASE-2022] - NPE in housekeeping kills RS
• [HBASE-2024] - [stargate] deletes not working as expected
• [HBASE-2025] - 0.20.2 accessed from older client throws UndeclaredThrowableException; frustrates rolling upgrade
• [HBASE-2026] - NPE in StoreScanner on compaction
• [HBASE-2027] - HConnectionManager.HBASE_INSTANCES leaks TableServers
• [HBASE-2033] - Shell scan 'limit' is off by one
• [HBASE-2034] - [Bulk load tools] loadtable.rb calls an undefined method `descendingIterator'
• [HBASE-2035] - Binary values are formatted wrong in shell
• [HBASE-2043] - Shell's scan broken
• [HBASE-2047] - Example command in the "Getting Started" documentation doesn't work
• [HBASE-2048] - Small inconsistency in the "Example API Usage"
• [HBASE-2049] - Cleanup HLog binary log output
• [HBASE-2052] - Make hbase more 'live' when comes to noticing table creation, splits, etc., for 0.20.3
• [HBASE-2054] - memstore size 0 is >= than blocking -2.0g size
• [HBASE-2057] - Cluster won't stop
• [HBASE-2060] - Missing closing tag in mapreduce package-info.java
• [HBASE-2062] - Metrics documentation outdated
• [HBASE-2068] - MetricsRate is missing "registry" parameter
• [HBASE-2075] - HBaseMaster requires HDFS superuser privileges due to waitOnSafeMode
• [HBASE-2082] - TableInputFormat is ignoring input scan's stop row setting
• [HBASE-2093] - [stargate] RowSpec parse bug
• [HBASE-2094] - hbase-2037 breaks mapreduce jobs going from 0.20.2 to 0.20.3
• [HBASE-2097] - Deadlock between HRegion.put and HRegion.close
• [HBASE-2101] - KeyValueSortReducer collapses all values to last passed
• [HBASE-2112] - New 'indexed' contrib is missing commons-lang.jar when packaged
• [HBASE-2113] - For indexed contrib, fast-forward to next row if no more results left... big performance improvement
• [HBASE-2119] - Fix top-level NOTICES.txt file. Its stale.
• [HBASE-2120] - [Stargate] Unable to delete column families
• [HBASE-2122] - [Stargate] Initializing scanners column families doesn't work
• [HBASE-2123] - Remove 'master' command-line option from PE.
• [HBASE-2127] - randomWrite mode of PerformanceEvaluation benchmark program writes only to a small range of keys
• [HBASE-2138] - unknown metrics type
• [HBASE-2144] - [shell] Now does \x20 for spaces
• [HBASE-2146] - RPC related metrics are missing in 0.20.3 since recent changes
• [HBASE-2149] - hbase.regionserver.global.memstore.lowerLimit is too low
• [HBASE-2154] - Fix Client#next(int) javadoc
• [HBASE-2156] - HBASE-2037 broke Scan
• [HBASE-2157] - LATEST_TIMESTAMP not replaced by current timestamp in KeyValue
• [HBASE-2160] - Can't put with ts in shell

Improvement

• [HBASE-1970] - Export does one version only; make it configurable how many it does
• [HBASE-1975] - SingleColumnValueFilter: Add ability to match the value of previous versions of the specified column
• [HBASE-2011] - Add zktop like output to HBase's master UI (zk.jsp)
• [HBASE-2021] - Add compaction details to master UI
• [HBASE-2029] - Reduce shell exception dump on console
• [HBASE-2031] - When starting HQuorumPeer, try to match on more than 1 address
• [HBASE-2045] - Update trunk and branch zk to just-release 3.2.2.
• [HBASE-2046] - Remove CompactionLimitThread on 0.20 branch
• [HBASE-2074] - Improvements to the hadoop-config script
• [HBASE-2076] - Many javadoc warnings
• [HBASE-2078] - Add JMX settings as commented out lines to hbase-env.sh
• [HBASE-2081] - Set the retries higher in shell since client pause is lower
• [HBASE-2095] - TableInputFormat should support more confs for the scanner
• [HBASE-2124] - Useless exception in HMaster on startup
• [HBASE-2133] - Increase default number of client handlers
• [HBASE-2152] - Add default jmxremote.{access|password} files into conf

New Feature

• [HBASE-1987] - The Put object has no simple read methods for checking what has already been added.
• [HBASE-2037] - Alternate indexed hbase implementation; speeds scans by adding indexes to regions rather secondary tables

The text area below allows the project release notes to be edited and copied to another document.