Welcome to the second edition of the Kudu Weekly Update. As with last week’s inaugural post, we’ll cover ongoing development and news in the Apache Kudu project on a weekly basis.
If you find this post useful, please let us know by emailing the kudu-user mailing list or tweeting at @ApacheKudu. Similarly, if you’re aware of some Kudu news we missed, let us know so we can cover it in a future post.
Development discussions and code in progress
-
Binglin Chang’s patch for per-tablet write quotas, described in last week’s post, was committed this week. This new experimental feature will be available in the upcoming 0.8.0 release.
-
Development discussion around improving the stability of the high-availability Kudu master continued on the design doc review.
-
Dan Burkert has been working on a design doc for a “scan tokens” API as described in the KUDU-1312 JIRA. He also posted a preliminary patch series for the Java implementation of the API.
This work will make it easier to integrate query and execution engines such as Spark, MapReduce, and Drill with Kudu. In particular, the API aims to support features such as partition pruning based on Kudu’s distribution schemas. See the design document for more details.
-
Ara Ebrahimi has been working on a Kudu sink for Flume. This will make it easier for users of Flume to build streaming ingest pipelines into Kudu tables. The code review is making steady progress with reviews by Mike Percy and Harsh J.
-
Jean-Daniel Cryans was the release manager for the past couple of releases and is acting as RM again for the upcoming 0.8.0 release. This week he reminded developers that the 0.8.0 train will soon be departing the station. JD is trying to follow the “train” release model: if a feature is ready when the release’s time comes, it will be included. Otherwise, it will have to wait for the next release to depart.
The thread hasn’t generated much discussion, as it’s just a reminder of the schedule that JD proposed a few months back. However, the focus this coming week will likely be on fixing blockers and getting the last bits and pieces of in-flight features in before a release candidate is cut.
Upcoming talks and meetups
- This week, O’Reilly and Cloudera will be hosting the Strata/Hadoop World in San Jose. The conference will feature two talks on Kudu:
- Thu, Mar 31, 2016. Strata Hadoop World. San Jose, CA, USA.
Fast data made easy with Apache Kafka and Apache Kudu (incubating). Presented by Ted Malaska and Jeff Holoman. - Wed, Mar 30, 2016. Strata Hadoop World. San Jose, CA, USA.
Hadoop’s storage gap: Resolving transactional-access and analytic-performance tradeoffs with Apache Kudu (incubating). Presented by Todd Lipcon.
Datanami has posted a preview of Todd Lipcon’s talk.