hbase-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Enis Söztutar <e...@apache.org>
Subject [VOTE] Third release candidate for HBase 1.0.0 (RC2) is available. Please vote by Feb 08 2015
Date Wed, 28 Jan 2015 04:30:14 GMT
It gives me great pleasure to announce that the third release candidate for
the release
1.0.0 (HBase-1.0.0RC2), is available for download at

Maven artifacts are also available in the temporary repository

Signed with my code signing key E964B5FF. Can be found here:

 Signed tag in the repository can be found here:

HBase 1.0.0 is the next stable release, and the start of "semantic
releases (See [1]).

The theme of 1.0.0 release is to become a stable base for future 1.x series
of releases. We aim to achieve at least the same level of stability of 0.98

1.0.0 release contains 162 fixes on top of 0.99.2 release. Together with the
previous 0.99.x releases, major changes in 1.0.0 are listed (but not
limited to)
below. Note that all previous 0.99.x releases are developer preview
releases, and will
NOT be supported in any form.

API Cleanup and changes
  1.0.0 introduces new APIs, and deprecates some of commonly-used
  client side APIs (HTableInterface, HTable and HBaseAdmin).
  We advise to update your application to use the new style of APIs, since
  deprecated APIs might be removed in future releases (2.x). See [2] and [3]
  for an overview of changes. All Client side API's are marked with
  InterfaceAudience.Public class, indicating that the class/method is an
  official "client API" for HBase. All 1.x releases are planned to be API
  compatible for these classes. See [1] for an overview.

Master runs a Region Server as well
  Starting with 1.0.0, the HBase master server and backup master servers
  also act as a region server. RPC port and info port for web UI is shared
  the master and region server roles. Active master can host regions of
  defined tables if configured (disabled by default). Backup masters will
  host regions.

Read availability using timeline consistent region replicas
  This release contains Phase 1 items for experimental "Read availability
  timeline consistent region replicas" feature. A region can be hosted in
  multiple region servers in read-only mode. One of the replicas for the
  will be primary, accepting writes, and other replicas will be sharing the
  data files. Read requests can be done against any replica for the region
  backup RPCs for high availability with timeline consistency guarantees.
  information can be found at HBASE-10070.

Online config change and other forward ports from 0.89-fb branch
  HBASE-12147 forward ported online config change which enables some of the
  configuration from the server to be reloaded without restarting the region

Other notable improvements in 1.0.0 (including previous 0.99.x) are
 - A new web skin in time for 1.0 (http://hbase.apache.org)
 - Automatic tuning of global memstore and block cache sizes
 - Various security, tags and visibility labels improvements
 - Bucket cache improvements (usability and compressed data blocks)
 - A new pluggable replication endpoint to plug in to HBase's inter-cluster
   replication to replicate to a custom data store
 - A Dockerfile to easily build and run HBase from source
 - Truncate table command
 - Region assignment to use hbase:meta table instead of zookeeper for faster
   region assignment (disabled by default)
 - Extensive documentation improvements
 - [HBASE-12511] - namespace permissions - add support from table creation
privilege in a namespace 'C'
 - [HBASE-12568] - Adopt Semantic Versioning and document it in the book
 - [HBASE-12640] - Add Thrift-over-HTTPS and doAs support for Thrift Server
 - [HBASE-12651] - Backport HBASE-12559 'Provide LoadBalancer with online
configuration capability' to branch-1
 - [HBASE-10560] - Per cell TTLs
 - [HBASE-11997] - CopyTable with bulkload
 - [HBASE-11990] - Make setting the start and stop row for a specific
prefix easier
 - [HBASE-12220] - Add hedgedReads and hedgedReadWins metrics
 - [HBASE-12090] - Bytes: more Unsafe, more Faster
 - [HBASE-12032] - Script to stop regionservers via RPC
 - [HBASE-11907] - Use the joni byte[] regex engine in place of j.u.regex
in RegexStringComparator
 - [HBASE-11796] - Add client support for atomic checkAndMutate
 - [HBASE-11804] - Raise default heap size if unspecified
 - [HBASE-11890] - HBase REST Client is hard coded to http protocol
 - [HBASE-12126] - Region server coprocessor endpoint
 - [HBASE-12183] - FuzzyRowFilter doesn't support reverse scans
 - [HBASE-12075] - Preemptive Fast Fail
 - [HBASE-12354] - Update dependencies in time for 1.0 release
 - [HBASE-12363] - Improve how KEEP_DELETED_CELLS works with MIN_VERSIONS
 - [HBASE-12434] - Add a command to compact all the regions in a
 - [HBASE-8707]  - Add LongComparator for filter
 - [HBASE-12286] - [shell] Add server/cluster online load of configuration
 - [HBASE-12361] - Show data locality of region in table page
  - [HBASE-12496] - A blockedRequestsCount metric
 - [HBASE-12730] - Backport HBASE-5162 (Basic client pushback mechanism) to
 - [HBASE-12731] - Heap occupancy based client pushback
 - [HBASE-12728] - buffered writes substantially less useful after removal
of HTablePool
 - [HBASE-5699] - Run with > 1 WAL in HRegionServer
  - Various fixes to REST server
 - Internal refactoring for abstracting away zookeeper usage
 - Better support for Cell interface internally in read and write paths for
   better performance and flexibility
 - Combining internal mvcc and seqId infrastructure
 - Ring buffer based WAL improvements
 - Switch to using new style of client APIs internally (in a lot of places)
 - Improvements in visibility labels
 - Perf improvements
 - Numerous improvements in other areas and bug fixes.

1.0.0 release has these changes in default behavior (see [4])
 - Zookeeper 3.4.x is required
 - Hadoop-2.x is required
 - JDK-1.7 is required
 - Default ports have changed from 600XX range into out of ephemeral 160XX
 - Hfile version 3 has been enabled by default.
 - Slab cache has been removed (use bucket cache instead)
 - Default heap for region servers (and master) is changed from 1GB to
   the jdk's default. On most platforms, it will default to 1/4 of physical
 - Some client facing APIs (HTableInterface, etc) has been deprecated and
 - Custom Filter implementations should switch to using the new recommended
 - Some authorization permissions (especially regarding namespaces) have
   changed (HBASE-12511, HBASE-8015)
 - Mapred package and classes under there has been un-deprecated.
  - Two new modules (jar files): hbase-annotations and hbase-rest.
 - HTrace dependency is moved to the Apache artifacts (
 - Various updates to dependency versions

Full list of the issues can be found at:
 - 1.0.0 issues  :
 - 0.99.2 issues :
 - 0.99.1 issues :
 - 0.99.0 issues :

Source Compatibility:
Client side code in HBase-1.0.0 is (mostly) source compatible with earlier
versions. Some minor API changes might be needed from the client side.

Wire Compatibility:
HBase-1.0.0 release is wire compatible with 0.98.x releases. Clients and
servers running in different versions as long as new features are not used
should be possible.
A rolling upgrade from 0.98.x clusters to 1.0.0 is supported as well. 1.0.0
introduces a new file format (hfile v3) that is enabled by default that
0.96.x code cannot read. Thus, rolling upgrade from 0.96 directly to 1.0.0
not supported.
1.0.0 is NOT wire compatible with earlier releases (0.94, etc).

Binary Compatibility:
Binary compatibility at the Java API layer with earlier versions (0.98.x,
0.96.x and 0.94.x) is not supported. You may have to recompile your client
code and any server side code (coprocessors, filters etc) referring to
hbase jars.

See [4] for upgrade instructions and extended discussion on the changes.

>From 0.98.x : Upgrade from 0.98.x in regular upgrade or rolling upgrade
is supported.

>From 0.96.x : Upgrade from 0.96.x is supported with a shutdown and restart
the cluster.

>From 0.94.x : Upgrade from 0.94.x is supported similar to upgrade from
0.94 -> 0.96. The upgrade script should be run to rewrite cluster level
See [5] for details.

Supported Hadoop versions
1.0.0 release drops support for Hadoop-1.x releases. Only Hadoop-2.x
releases are supported. Hadoop-2.4.x, Hadoop-2.5.x and Hadoop-2.6.x releases
are the most tested hadoop releases and we recommend running with those
(or later versions). Earlier Hadoop-2 based releases (hadoop-2.2.x and
are not tested to the full extend. More information can be found here:

Supported Java versions
1.0.0 release drops support for JDK6. Only JDK7 is supported. JDK8 support
is experimental. More information can be found here:

Please try to test and vote on this release by Feb 08 2015 11:59PM PDT.
Due to the importance of 1.0 release we will have a 10 day voting period
to cover sufficient testing. Please notify here if you need more time to
test the RC.

[] +1 Release the artifacts as 1.0.0
[] -1 DO NOT release the artifacts as 1.0.0, because...

Last but not least, 1.0.0 release has been in a long time making with
contributions from a very large group of awesome people and hard work from
committers and contributors. We would like to extend our thanks to everybody
who worked on this release or contributed to HBase over the years.

[1] https://hbase.apache.org/book/upgrading.html#hbase.versioning
[2] http://www.slideshare.net/xefyr/apache-hbase-10-release
[3] http://s.apache.org/hbase-1.0-api
[4] https://hbase.apache.org/book/upgrade1.0.html#upgrade1.0.changes
[5] https://hbase.apache.org/book/upgrade1.0.html#upgrade1.0.from.0.94


  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message