Here is another blocker for running mapreduce on top hadoop 2.0.3-alpha: https://issues.apache.org/jira/browse/YARN-449
On Tue, Feb 26, 2013 at 12:03 PM, Ted Yu <[email protected]> wrote: > https://issues.apache.org/jira/browse/YARN-429 was logged for this issue. > > FYI > > > On Fri, Feb 22, 2013 at 1:12 PM, Andrew Purtell <[email protected]>wrote: > >> I agree, thanks a lot for doing this. >> >> >> On Fri, Feb 22, 2013 at 1:07 PM, Ted Yu <[email protected]> wrote: >> >> > bq. to nail down the actual issue will help making our case. >> > >> > That is what I am trying to do. Will ping Arun at the meetup this >> afternoon >> > about when the next release is coming. >> > >> > Cheers >> > >> > On Fri, Feb 22, 2013 at 1:02 PM, Stack <[email protected]> wrote: >> > >> > > Thanks for trying to fix it Ted. I think we need to carry our >> difficulty >> > > back up to our parent project brethren (Sounds like you have done this >> > > off-list; lets do it in jira/mail too). Your digging in to nail down >> the >> > > actual issue will help making our case. Maybe we can get/help the >> > > upstream crew make another release (I don't see a note on a new 2.0 >> > release >> > > up on hadoop -- have I missed it)? >> > > >> > > Thanks, >> > > St.Ack >> > > >> > > >> > > On Fri, Feb 22, 2013 at 11:49 AM, Ted Yu <[email protected]> wrote: >> > > >> > > > Trying out #2, I got: >> > > > >> > > > >> > > > >> > > >> > >> testExcludeAllFromMinorCompaction(org.apache.hadoop.hbase.mapreduce.TestHFileOutputFormat) >> > > > Time elapsed: 35.746 sec <<< FAILURE! >> > > > java.lang.AssertionError >> > > > at org.junit.Assert.fail(Assert.java:86) >> > > > at org.junit.Assert.assertTrue(Assert.java:41) >> > > > at org.junit.Assert.assertTrue(Assert.java:52) >> > > > at >> > > > >> > > > >> > > >> > >> org.apache.hadoop.hbase.mapreduce.TestHFileOutputFormat.runIncrementalPELoad(TestHFileOutputFormat.java:468) >> > > > at >> > > > >> > > > >> > > >> > >> org.apache.hadoop.hbase.mapreduce.TestHFileOutputFormat.testExcludeAllFromMinorCompaction(TestHFileOutputFormat.java:713) >> > > > >> > > > Arun Murthy got to know this defect. Another 2.0 release is coming >> soon >> > > > (next week, hopefully). >> > > > >> > > > I plan to wait for the new release. >> > > > >> > > > Thanks >> > > > >> > > > On Fri, Feb 22, 2013 at 11:39 AM, Ted Yu <[email protected]> >> wrote: >> > > > >> > > > > The root cause was the removal of capacity-scheduler.xml< >> > > > >> > > >> > >> http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/conf/capacity-scheduler.xml >> > > > >> > > > from >> > > > > 2.0.3-alpha maven artifacts. >> > > > > This config file is needed to start CapacityScheduler which is the >> > > > default. >> > > > > >> > > > > After talking to Sid, I got two suggestions: >> > > > > 1. embed minimal capacity-scheduler.xml< >> > > > >> > > >> > >> http://svn.apache.org/viewvc/hadoop/common/branches/branch-2/hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-resourcemanager/conf/capacity-scheduler.xml >> > > > >> > > > in >> > > > > our src/test/resources >> > > > > 2. use FiFo scheduler instead which doesn't require presence of >> > config >> > > > file >> > > > > >> > > > > I think #2, though temporary fix, is easy to change once a future >> 2.0 >> > > > > release addresses this issue. >> > > > > >> > > > > Once consensus is reached, I will attach another patch to >> HBASE-7904 >> > > > > >> > > > > Cheers >> > > > > >> > > > > >> > > > > On Fri, Feb 22, 2013 at 10:31 AM, Andrew Purtell < >> > [email protected] >> > > > >wrote: >> > > > > >> > > > >> For minicluster unit tests? In my opinion, no. >> > > > >> >> > > > >> >> > > > >> On Fri, Feb 22, 2013 at 10:10 AM, Ted Yu <[email protected]> >> > wrote: >> > > > >> >> > > > >> > I can try to find the proper usage. >> > > > >> > >> > > > >> > The question is: do we have to maintain compatibility with >> > > > 2.0.2-alpha ? >> > > > >> > >> > > > >> > Thanks >> > > > >> > >> > > > >> > On Fri, Feb 22, 2013 at 10:07 AM, Andrew Purtell < >> > > [email protected] >> > > > >> > >wrote: >> > > > >> > >> > > > >> > > Sigh. Every release seems to break our uses of MiniMRCluster >> all >> > > > over >> > > > >> > > again. >> > > > >> > > >> > > > >> > > >> > > > >> > > On Fri, Feb 22, 2013 at 10:04 AM, Ted Yu < >> [email protected]> >> > > > wrote: >> > > > >> > > >> > > > >> > > > I temporarily reverted this change due to: >> > > > >> > > > >> > > > >> > > > org.apache.hadoop.yarn.YarnException: Failed to Start >> > > > >> > > > org.apache.hadoop.mapred.MiniMRCluster >> > > > >> > > > at >> > > > >> > > > >> > > > >> > > > >> > > > >> > > >> > > > >> > >> > > > >> >> > > > >> > > >> > >> org.apache.hadoop.yarn.service.CompositeService.start(CompositeService.java:78) >> > > > >> > > > at >> > > > >> > > > >> > > > >> > > > >> > > > >> > > >> > > > >> > >> > > > >> >> > > > >> > > >> > >> org.apache.hadoop.mapred.MiniMRClientClusterFactory.create(MiniMRClientClusterFactory.java:75) >> > > > >> > > > at >> > > > >> > > > >> > > > >> >> > org.apache.hadoop.mapred.MiniMRCluster.<init>(MiniMRCluster.java:180) >> > > > >> > > > at >> > > > >> > > > >> > > > >> >> > org.apache.hadoop.mapred.MiniMRCluster.<init>(MiniMRCluster.java:170) >> > > > >> > > > at >> > > > >> > > > >> > > > >> >> > org.apache.hadoop.mapred.MiniMRCluster.<init>(MiniMRCluster.java:162) >> > > > >> > > > at >> > > > >> > > > >> > > > >> >> > org.apache.hadoop.mapred.MiniMRCluster.<init>(MiniMRCluster.java:123) >> > > > >> > > > at >> > > > >> > > > >> > > > >> > > > >> > > > >> > > >> > > > >> > >> > > > >> >> > > > >> > > >> > >> org.apache.hadoop.hbase.HBaseTestingUtility.startMiniMapReduceCluster(HBaseTestingUtility.java:1574) >> > > > >> > > > ... >> > > > >> > > > Caused by: org.apache.hadoop.yarn.YarnException: >> > > > >> > > > java.lang.IllegalStateException: Queue configuration >> missing >> > > child >> > > > >> > queue >> > > > >> > > > names for root >> > > > >> > > > at >> > > > >> > > > >> > > > >> > > > >> > > > >> > > >> > > > >> > >> > > > >> >> > > > >> > > >> > >> org.apache.hadoop.yarn.server.MiniYARNCluster$ResourceManagerWrapper.start(MiniYARNCluster.java:181) >> > > > >> > > > at >> > > > >> > > > >> > > > >> > > > >> > > > >> > > >> > > > >> > >> > > > >> >> > > > >> > > >> > >> org.apache.hadoop.yarn.service.CompositeService.start(CompositeService.java:68) >> > > > >> > > > ... 38 more >> > > > >> > > > Caused by: java.lang.IllegalStateException: Queue >> > configuration >> > > > >> missing >> > > > >> > > > child queue names for root >> > > > >> > > > at >> > > > >> > > > >> > > > >> > > > >> > > > >> > > >> > > > >> > >> > > > >> >> > > > >> > > >> > >> org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.CapacityScheduler.parseQueue(CapacityScheduler.java:328) >> > > > >> > > > at >> > > > >> > > > >> > > > >> > > > >> > > > >> > > >> > > > >> > >> > > > >> >> > > > >> > > >> > >> org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.CapacityScheduler.initializeQueues(CapacityScheduler.java:255) >> > > > >> > > > at >> > > > >> > > > >> > > > >> > > > >> > > > >> > > >> > > > >> > >> > > > >> >> > > > >> > > >> > >> org.apache.hadoop.yarn.server.resourcemanager.scheduler.capacity.CapacityScheduler.reinitialize(CapacityScheduler.java:220) >> > > > >> > > > at >> > > > >> > > > >> > > > >> > > > >> > > > >> > > >> > > > >> > >> > > > >> >> > > > >> > > >> > >> org.apache.hadoop.yarn.server.resourcemanager.ResourceManager.init(ResourceManager.java:226) >> > > > >> > > > at >> > > > >> > > > >> > > > >> > > > >> > > > >> > > >> > > > >> > >> > > > >> >> > > > >> > > >> > >> org.apache.hadoop.yarn.server.MiniYARNCluster$ResourceManagerWrapper.start(MiniYARNCluster.java:161) >> > > > >> > > > >> > > > >> > > > Cheers >> > > > >> > > > >> > > > >> > > > On Fri, Feb 22, 2013 at 9:55 AM, Andrew Purtell < >> > > > >> [email protected]> >> > > > >> > > > wrote: >> > > > >> > > > >> > > > >> > > > > Have you tried running the mapreduce unit tests with this >> > > > version >> > > > >> > Ted? >> > > > >> > > > Just >> > > > >> > > > > curious. They have given us grief in the past. >> > > > >> > > > > >> > > > >> > > > > >> > > > >> > > > > On Fri, Feb 22, 2013 at 9:35 AM, Ted Yu < >> > [email protected]> >> > > > >> wrote: >> > > > >> > > > > >> > > > >> > > > > > Hi, >> > > > >> > > > > > I have integrated the following JIRA to trunk: >> > > > >> > > > > > HBASE-7904 Upgrade hadoop 2.0 dependency to 2.0.3-alpha >> > > > >> > > > > > >> > > > >> > > > > > Earlier this week, Ramkrishna encountered some problem >> > where >> > > > >> HBase >> > > > >> > > was >> > > > >> > > > > > built against 2.0.2-alpha but hadoop cluster deployment >> > was >> > > > >> using >> > > > >> > > > > > 2.0.3-alpha >> > > > >> > > > > > >> > > > >> > > > > > In case you use HBase against hadoop 2.0, please note >> this >> > > > >> change. >> > > > >> > > > > > >> > > > >> > > > > > Cheers >> > > > >> > > > > > >> > > > >> > > > > >> > > > >> > > > > >> > > > >> > > > > >> > > > >> > > > > -- >> > > > >> > > > > Best regards, >> > > > >> > > > > >> > > > >> > > > > - Andy >> > > > >> > > > > >> > > > >> > > > > Problems worthy of attack prove their worth by hitting >> > back. - >> > > > >> Piet >> > > > >> > > Hein >> > > > >> > > > > (via Tom White) >> > > > >> > > > > >> > > > >> > > > >> > > > >> > > >> > > > >> > > >> > > > >> > > >> > > > >> > > -- >> > > > >> > > Best regards, >> > > > >> > > >> > > > >> > > - Andy >> > > > >> > > >> > > > >> > > Problems worthy of attack prove their worth by hitting back. >> - >> > > Piet >> > > > >> Hein >> > > > >> > > (via Tom White) >> > > > >> > > >> > > > >> > >> > > > >> >> > > > >> >> > > > >> >> > > > >> -- >> > > > >> Best regards, >> > > > >> >> > > > >> - Andy >> > > > >> >> > > > >> Problems worthy of attack prove their worth by hitting back. - >> Piet >> > > Hein >> > > > >> (via Tom White) >> > > > >> >> > > > > >> > > > > >> > > > >> > > >> > >> >> >> >> -- >> Best regards, >> >> - Andy >> >> Problems worthy of attack prove their worth by hitting back. - Piet Hein >> (via Tom White) >> > >
