I requested Xuan Gong to help me dig into this and he figured out the test-failures.
The patch for MAPREDUCE-5006 addresses the test-issues. I just merged MAPREDUCE-5006 into branch-2.0.4-alpha.
Will report today's nightly build's result. Once it is a success, we are good for the release.
+Vinod Kumar Vavilapalli
On Apr 4, 2013, at 3:18 PM, Konstantin Boudnik wrote:
> Unless the failing unit tests are considering blockers...
> On Thu, Apr 04, 2013 at 01:22PM, Vinod Kumar Vavilapalli wrote:
>> Great news! We can go forward with releasing 2.0.4-alpha now, thanks!
>> +Vinod Kumar Vavilapalli
>> On Apr 3, 2013, at 3:29 PM, Roman Shaposhnik wrote:
>>> Thanks to the tireless work of the Bigtop and Hadoop committers
>>> we've made quite a bit of progress on 2.0.4-alpha. The build
>>> pipeline that we've setup seems to be rather nice (and perhaps
>>> could be utilized for all future releases of Hadoop). The first
>>> stage of the pipeline is available here:
>>> and the second one over here:
>>> Anybody who's interested in helping with a release is more than
>>> welcome to get either tarballs or packages from these 2 jobs.
>>> As far as testing is concerned we're now completely unblocked
>>> for both unsecure and fully kerberized testing modes and Bigtop
>>> is proceeding with the rest of the testing process. We've got
>>> 4 unit test failures left:
>>> which I'm going to file JIRAs for.
>>> Given that I don't see any current showstoppers in ASF JIRA
>>> and hoping that the rest of Bigtop testing doesn't discover
>>> anything major do you guys think we can release sometime
>>> next week?