On 30 Jun 2017, at 03:40, Andrew Wang <andrew.w...@cloudera.com<mailto:andrew.w...@cloudera.com>> wrote:
Hi all, As always, thanks to the many, many contributors who helped with this release! I've prepared an RC0 for 3.0.0-alpha4: http://home.apache.org/~wang/3.0.0-alpha4-RC0/ whats the git commit ID? or signed tag? The standard 5-day vote would run until midnight on Tuesday, July 4th. Given that July 4th is a holiday in the US, I expect this vote might have to be extended, but I'd like to close the vote relatively soon after. I've done my traditional testing of a pseudo-distributed cluster with a single task pi job, which was successful. Normally my testing would end there, but I'm slightly more confident this time. At Cloudera, we've successfully packaged and deployed a snapshot from a few days ago, and run basic smoke tests. Some bugs found from this include HDFS-11956, which fixes backwards compat with Hadoop 2 clients, and the revert of HDFS-11696, which broke NN QJM HA setup. Vijay is working on a test run with a fuller test suite (the results of which we can hopefully post soon). My +1 to start, Best, Andrew I haven't tested with this alpha, but spark 2.2+ can't read files on Azure: https://issues.apache.org/jira/browse/HADOOP-14598 Cause: https://issues.apache.org/jira/browse/HADOOP-14383 I can see how to do a test we can put into hadoop-common (for trunk/2.9); if the bug exists in this RC then I think we'll need to revert the HADOOP-14598 patch for now & then work on a fix Sorry —only discovered it this week...an unexpected combination of features which only become visible when you take the latest bits of two projects and then run some blobstore tests downstream (https://github.com/hortonworks-spark/cloud-integration/tree/master/cloud-examples/)