Version 2.1.0 is a feature release.
New features:
- Instrumentation: if the
instrumentable
gem is loaded, the client exposes several event hooks toActiveSupport::Notifications
. Read the README for more information, and if you'd like other events to be instrumented, please file GitHub issues. Instrumentation was developed by Ryan Daigle. - CRDTs support the
returnbody
option, and use it by default. This means that unless specified otherwise, CRDTs will update themselves on a write.
Small changes:
- UTF-8 support is now tested against.
- RSpec 3.1 is now supported, although RSpec 3.0 still works.
- Specs no longer use gratuitous "should"s.
Version 2.0.0 is a major new version with many new features, API changes, and feature removals.
New features:
- Yokozuna: full-text search built on Solr and powered by Riak.
- Riak security: TLS-encrypted and authenticated protocol buffers, access control, and more!
- Convergent Replicated Data Types (CRDTs): counters, maps, and sets, all with convenient and safe distributed semantics.
- Bucket types: the building blocks of Yokozuna, access control, and CRDTs.
API changes:
- Exceptions raised by the client are subclasses of
Riak::Error
. - The internals of the Beefcake-based protocol buffers support have been refactored for reliability and maintainability.
- The Beefcake version has been bumped to 1.0 for improvements in speed and memory usage.
- Tests now use RSpec 3.
Removed:
- HTTP support has been removed from the Riak Ruby Client in favor of focusing on Protocol Buffers.
- The included test-server has been removed. Tests now require a Riak node to be configured and run independently of the test suite.
Release 1.4.2 fixes a couple bugs.
Bugfixes:
- 2i Requests over PBC block forever when 0 results match in 1.4.x, reported by Sean "graphex" McKibben in basho#121 and basho#122
- RObject#links is an Array when loaded from PBC, reported by Dan Pisarski in basho#123
Release 1.4.1 fixes a few minor bugs and issues.
Issues:
- Test for object existence using head request, reported and fixed by Elias "eliaslevy" Levy in basho#102
Bugfixes:
- License missing from gemspec, reported by Benjamin "bf4" Fleischer in basho#108
- Debugger required by Gemfile, reported by Basho Giddyup in basho#114
- Issue when reading Git-based version numbers, reported and fixed by jacepp in basho#120
Release 1.4.0 adds support for Riak 1.4 and fixes a few bugs.
Features for all Riak versions:
- Multi-get parallelizes fetching multiple objects from one or more buckets.
Features for Riak 1.4 and newer:
- Bucket properties are settable and resettable over Protocol Buffers.
- Distributed counters are implemented by the
Riak::Counter
class. - The full set of improvements to Secondary Indexes
are available, including pagination, streaming, and return_terms.
These features are available through the existing
Bucket#get_index
interface as well as the newRiak::SecondaryIndex
interface. - The new streaming bucket list is available in the Ruby client.
- Setting timeout values for object CRUD, key listing, and bucket listing is now possible.
Bugfixes:
- Tests pass and don't stall in Ruby 2.0.
- Zero-length key and bucket names are forbidden in the client.
- Test server works with Riak 1.4.
Release 1.2.0 adds support for Riak 1.3 and fixes a number of bugs.
Features:
- The "clear bucket properties" feature has been added. This resets modified bucket properties to the defaults on Riak 1.3+ clusters.
- Anonymous "strfun" MapReduce functions written in Erlang can now be sent from the client, if they are enabled on the server-side.
Bugfixes:
- The WalkSpec class now properly includes the Translation module.
- The Protocol Buffers transport now extracts the bucket name before submitting secondary index queries.
- Search query results returned over PBC are assumed to be UTF-8 encoded.
- The newer Excon API is now supported (>= 0.19.0).
- When enabling the search commit hook, the 'precommit' property will now be checked more safely.
Release 1.1.1 fixes a minor bug with Net::HTTP on Ruby 1.8.7 with patch level less than 315, where an exception would cause closing the socket before it was opened.
Release 1.1.0 includes full Riak 1.2 compatibility, and includes improvements to the handling of siblings, the node generation tools, and resolves a number of important bugs.
Features:
- Client features are enabled or disabled based on the detected Riak version.
- Riak 1.2 compatibility, including search and 2I over Protocol Buffers.
- Phaseless MapReduce (which was available in 1.1) is allowed, using feature detection to determine whether an exception is raised.
- Conditional store_object operations on Protocol Buffers use the message features available since Riak 1.0.
- The integration test-suite can be run without generating a test node, which lets us support riak_test.
Bugfixes:
- URL-escaping now allows some normally URI-safe characters to be escaped.
- JRuby should be more reliable when attaching to a generated node's console.
- The client backend pool has been extracted to the Innertube gem, which is now a dependency.
- Fix a documentation issue around key-filters.
- Fix RSpec formatter and deprecation errors.
- Object siblings are now a separate class (RContent) rather than
being unclean copies of the parent RObject. If only one sibling
exists, the original accessors (e.g.
content_type
,data
) will behave as expected. When more than one sibling exists, they will raiseRiak::Conflict
. This should prevent unintentional storing of unresolved objects back into Riak asmultipart/mixed
values. Riak::Client#ssl=
won't blow away existingssl_options
if set totrue
.- Generated nodes will ensure that the source's
ssl_distribution.args_file
exists by invokingriak chkconfig
. - Copy fixes for the
$key
index on the memory/test backend from riak_kv. - The shape of MapReduce results will no longer be changed by the Protocol Buffers backend, which manifested as kept phases without results being removed from the return value. Implementing this required all HTTP requests to use streaming, even if invoked without a block.
Release 1.0.5 fixes a bug with the RubyGems packaging that
inadvertently included the pkg
directory, which might have included
old gem versions. No client functionality has changed with this
release.
Release 1.0.4 fixes some bugs and adds configurable timeouts to the Excon HTTP backend.
NOTE This will likely be the last release in the 1.0.x series. The planned changes for 1.1.x are:
Riak::Client::Pool
will be replaced by theinnertube
gem, which is its extraction.- Riak 1.2 will be fully supported, including the new native 2I and Search features over PBC.
- A richer exception hierarchy so that applications can deal more intelligently with request failures.
Changes in 1.0.4:
- A function in the
app_helper
module that does not exist on Riak 1.1 and earlier was copied into the KV test backend. - Excon's configuration logic was made more idempotent.
- Added timeout support to the Excon HTTP backend. [Mat Brown]
- Corrected an misnamed constant in Excon which would cause timeouts not to be recognized as network errors.
- The
Riak::TestServer
is now compatible with Riak 1.2. - A documentation error around
RObject#data
in the README was fixed. [dn@wortbit.de] - Fixed an ETS table leak in the testing backend.
- Deprecation warnings for later versions of MultiJson are now resolved.
Release 1.0.3 fixes some bugs and adds support for secondary indexes
when using Riak::TestServer
.
- Added tests for secondary index features to the unified backend examples.
- Added secondary index support to
riak_kv_test_backend
. Full support for this feature will be available viariak_kv_memory_backend
in the next major Riak release. See riak_kv #314. - The console log (
lager_console_backend
) is now enabled on generated nodes. Riak::Node::Console
no longer overrides theSIGWINCH
signal handler.- Excon versions >= 0.7.0 are now supported.
- IO-style objects will now be emitted properly when using the
NetHTTPBackend
. #1 - The Riak version filter for integration specs is now more correct.
Riak::RObject#url
has been removed because its accuracy cannot be maintained when connected to multiple Riak nodes or to Riak via PBC. #3- Index entries on
Riak::RObject
can be mass-overwritten usingRiak::RObject#indexes=
while maintaining the proper internal semantics. #17 - Nodes should now generate properly when the
riak
script is a symlink (e.g. Homebrew). #26
Release 1.0.2 relaxes the multi_json dependency so that the client will function with Rails 3.2. Version 1.0.1 was yanked.
Release 1.0.1 is a minor bugfix/patch release. Included in this release are:
- I18n messages now include the French locale. [Eric Cestari]
- SSL configuration should work again. [Adam Hunter]
- The version comparison when checking Excon compatibility should now handle large version numbers correctly. [Srdjan Pejic]
- There is now a spec to verify that the
riak_kv
add_paths
setting is not clobbered by theRiak::TestServer
when adding the location of the test backend code.
Release 1.0.0 is a major feature release and is the first where
riak-client
, ripple
, and riak-sessions
will be released
independently (see below). Because there too many individual changes
to recount, this entry will cover the major features and bugfixes
present in the release.
riak-client
is fully compatible with Riak 1.0.x and
(yet-to-be-released) 1.1.x, including supporting secondary indexes,
integrated search, and cluster membership commands.
Riak::Client
can now connect to multiple Riak nodes at once. This
greatly improves throughput and allows the client to recover from
intermittent connection errors while continuing normal operation. To
enable this, all uses of the Pump/Fiber logic were removed in favor of
connection pools from which any new request can draw an existing or
create a new connection. Which node is selected for any new connection
is based on a quickly-decaying EWMA of its success rate on recent
requests. A huge thanks to Kyle Kingsbury
who did most of the work on this!
The Riak::TestServer
class has been generalized such that you can
generate regular nodes and even clusters that store data on disk. This
is especially useful if you want separate nodes or clusters for each
project that uses Riak, and to keep them separate from your base
install. TestServer
also now launches the node in a separate process
(not a child process) so you can keep it running between test
suites. Clearing the in-memory data is performed by connecting to the
console via the exposed Unix pipes, rather than over stdio.
An important part of dealing with eventual consistency is the ability
to handle when conflicts (also called siblings) are created. Now you
can resolve them automatically by registering blocks (callbacks) using
Riak::RObject.on_conflict
. The block will be called when fetching a
key in conflict and receives a RObject
that has siblings. To resolve
the conflict, it simply returns the resolved object, or nil if it
didn't handle the conflict. A huge thanks to
Myron Marston who implemented this!
Before, serialization of Ruby objects into Riak was constrained to
three formats: JSON, YAML and Marshal. Now you can define your own
serializers so that you can store data in BSON, MsgPack, NetStrings,
or whatever format you like. Use Riak::Serializers[content_type] = serializer
to assign a serializer for the selected media type. The
serializer must respond to #dump
and #load
. (More handiwork of Myron
Marston, thanks!)
If you don't like the keys that Riak hands out when you store an
RObject
without a key, and you want something naturally ordered, you
can now generate them client-side using Riak::Stamp
, which will
generate 64-bit integers in a fashion similar to Twitter's Snowflake,
but uses Riak::Client#client_id
as the machine identifier.
In an effort to decouple development of the individual projects and
reduce top-level dependencies, the ripple
repository was split into
new repositories containing its corresponding sub-projects.
Additionally, the Riak::CacheStore
has become its own project/gem.
The new gem and repository locations are below:
riak-client
— basho/riak-ruby-clientripple
— seancribbs/rippleriak-sessions
— seancribbs/riak-sessionsriak-cache
— seancribbs/riak-cache
Attempting to use the Protocol Buffers transport with a 0.14.x cluster may cause the connection to dump because of incompatibilities in certain protocol messages. This will be addressed in a future patch/bugfix release.
The new node generation and test server intermittently fails on JRuby, specifically from deadlocks related to blocking opens for the console FIFOs. The JRuby team has helped on this issue, but there may not be a clear resolution path until JRuby 1.7 or later.
Other known issues may be found on the Github issue tracker.