[riak] Added test unit.
[riak] Set eventual consistency as default in both README.md and riak.properties file, as strong consistency it's not the main model adopted by the db (check http://docs.basho.com/riak/kv/2.1.4/developing/app-guide/strong-consistency/ for further info).
[riak] Fixed getFields() function by transforming it into createResultHashMap(). This function creates the HashMap to give as result appropriately.
[riak] Removed scan transaction support for strong consistency model. 2i search is not allowed for this model, neither there is a suitable search mean to use to achieve a proper scan transaction.
[riak] Done minor fixes to code and comments.
[riak] Packaged into com.yahoo.ycsb.db.riak.
[riak] Removed unused dependencies.
[riak] Bugfix for update() function that returned positive even with no nodes in the cluster.
Eventually it would be nice to move this out into a proper generator with enums to
save space, but as users may have extended CoreWorkload and could be using the
"operationchooser" field, we'll just break it out into a static function so as to
avoid breaking anyone's code.
The binding is based on the one created by Basho Technologies (available at https://github.com/basho-labs/YCSB/tree/master/riak).
[riak] Made compatible with the latest YCSB version.
[riak] Added strong consistency test support and made it default.
[riak] Added missing results return based on provided fields in scan/read functions.
[riak] Added TIME_OUT status return code for those transactions which take too long to complete.
[riak] Added flexible test parameter specification via customizable configuration file and command line options support.
[riak] Added test unit.
[riak] Removed unused Constants.java file.
[riak] Removed many unused function in the RiakUtils.java file.
[riak] Choosed to use executeAsync methods to resolve a bug that didn't allow the benchmark to end whenever one or more nodes of the cluster are killed.
[riak] Fixed license issues in all files.
[riak] Done some refinements to the code.
This commit fixes an issue with the index creation logic when the
elasticsearch.newdb flag is set to true. Namely, when running in local
mode the index would never exist (since the node always starts with a
clean temporary directory) but with elasticsearch.newdb an attempt would
be made to delete the index which would cause the workload to fail.
This commit also does a cleanup of the index creation logic using an
explicit create index request and passing the index settings along as
part of that index creation.
nygard_89
Dani
Kevin Risden
Chrisjan Matser
yuyanting
Sean Busbey
Kevin Risden
Chris Larsen
Jason Tedor
stack
Michael Nitschinger