Hadoop 3 timeline

[Robert Kanter <rk...@cloudera.com>]

Hi all,

FYI, Andrew Wang (the release manager for Hadoop 3) is keeping updates on this
wiki page
<https://cwiki.apache.org/confluence/display/HADOOP/Hadoop+3.0.0+release>,
for those who are interested.

Alpha 2 was released about a week ago.  There's going to be an Alpha 3, and
then a Beta 1.  After that, we'll have GA, which is currently targeting
July.

We've seen two issues so far with Hadoop 3:

   1. Hadoop 3 no longer issues delegation tokens in non-secure clusters (
   YARN-4126 <https://issues.apache.org/jira/browse/YARN-4126>).  I believe
   this has been reverted in Alpha 2 via YARN-5882
   <https://issues.apache.org/jira/browse/YARN-5882>.
   2. Hadoop 3 no longer allows duplicate filenames in the distributed
   cache, even if they're the same file.  We had a problem like this long ago
   in Hadoop 2.0.2-alpha (see OOZIE-1089
   <https://issues.apache.org/jira/browse/OOZIE-1089>), but the workaround
   for that doesn't seem to work here.  I believe we're still hitting this
   issue.

I imagine that Oozie 5 will be the first release that will support Hadoop 2
and Hadoop 3, so we should keep an eye on the Hadoop 3 progress.


- Robert