Tested download with md5: 8f40198ed18bef28aeea1401ec536cb9 Tried to verify the GPG signature, but Chris is not in http://download.nextag.com/apache/hadoop/core/KEYS - he should be added there if he is going to sign releases.
I ran unit tests on my machine at home - TestStreamingExitStatus failed with an OOME. I think it's exactly MAPREDUCE-587. Aside from that, all unit tests passed. I also ran a few jobs on a pseudo-distributed cluster and it worked fine. Since this is just a test bug, and in contrib, I think we should release anyway. Meanwhile let's commit MAPREDUCE-587 to branch-20 before the next release. I'll reopen that JIRA. So, [non-binding] +1 from me. Thanks for the hard work, Chris. -Todd On Mon, Feb 22, 2010 at 11:52 PM, Dhruba Borthakur <dhr...@gmail.com> wrote: > +1. Looks good to me. Ran unit tests. > > -dhruba > > > On Fri, Feb 19, 2010 at 11:08 AM, Stack <st...@duboce.net> wrote: > >> +1 >> >> I put up on a small cluster under load. Seems to work fine. Trolled >> logs a while. Nothing out of the ordinary. Checked docs. They look >> grand. >> >> St.Ack >> >> On Fri, Feb 19, 2010 at 1:07 AM, Chris Douglas <cdoug...@apache.org> >> wrote: >> > There are now only two consistently failing testcases in my >> > environment, both in the capacity-scheduler contrib module: >> > >> > org.apache.hadoop.mapred.TestJobInitialization >> > org.apache.hadoop.mapred.TestQueueCapacities >> > >> > neither of which is a regression from 0.20.1 >> > >> > http://people.apache.org/~cdouglas/0.20.2-rc4 >> > >> > Please try it out. -C >> > >> > > > > -- > Connect to me at http://www.facebook.com/dhruba >