Over the next few weeks there will be a variety of contributions reaching prime-time status. One of them is the multi-operation transaction support for Zookeeper. This is the largest change to the client-side Zookeeper API that I know of since Zookeeper first went public and it enables new classes of applications to use Zookeeper in new ways. We also are supporting a fork in YCSB that adds data checking, performance and code improvements and soon will support clustered load generation. We are addressing smaller issues as well and will have minor fixes to contribute to Pig, Flume and Hive.
During our stealthy time we have worked closely with key HBase™ committers to qualify HBase™ on the MapR platform and that effort has uncovered several race conditions that are now fixed. We should have information to release soon doing on transactionally atomic snapshots in HBase™.
Then in June and July look for substantial contributions to Hadoop's core map-reduce structure that will result in major performance improvements.
We want to congratulate the individuals and corporations that have given to open source in the past that have resulted in the rapid, viral adoption of Hadoop. We’re excited to join the growing number of commercial entities and the Hadoop community to provide innovation and advancements.