Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
MapReduce, mail # dev - Heads up: branch-2.1-beta

Copy link to this message
Re: Heads up: branch-2.1-beta
Arun C Murthy 2013-06-16, 23:50
Responses inline:

On Jun 16, 2013, at 1:04 PM, Roman Shaposhnik wrote:

> But there's a bit of bad news too (or at least the news that need to
> be triaged). At
> this point I don't know whether the Hadoop code is to blame or the
> tests/components
> themselves -- all I know is that these tests passed with Hadoop 2.0.5-alpha:
>   1. HDFS append integration tests failed:
>            http://bigtop01.cloudera.org:8080/view/Bigtop-trunk/job/Bigtop-trunk-runsmokes/lastCompletedBuild/testReport/org.apache.bigtop.itest.hadoop.hdfs/TestFileAppend/testMultipleOutputStreamFailure/
>       It seems that there's some incompatibility between the client code
>       that was compiled against Hadoop 2.0.5 (as part of Bigtop 0.6.0 release)
>       and the current Hadoop 2.1.0.

This is well known, we need to recompile against hadoop-2.1.0-beta.

>    2. Quite a few Sqoop tests ended up failing because of what seems
> like AM not
>    realizing that one of the tasks exited and waiting for it to
> timeout. In the end
>    the task is getting killed like this:
>         AttemptID:attempt_1371348647940_0030_m_000000_2 Timed out
> after 600 secsContainer killed by the ApplicationMaster.
>     but it takes a VERY long time (on the task side the log is
> attached bellow).

Essentially, this is a Sqoop error - we could investigate why it took 600s, but doesn't look like a blocker to me.

We'll probably need another 2.1.1-beta anyway...

>     3. There's a couple of Hive tests (out of more than a dozen) that
> failed in a pretty odd way
>          (scroll to the very bottom of every page to see the excpetion):
>             http://bigtop01.cloudera.org:8080/view/Bigtop-trunk/job/Bigtop-trunk-runsmokes/59/testReport/org.apache.bigtop.itest.hivesmoke/TestHiveSmokeBulk/testHiveBulk_auto_join20_/
>             http://bigtop01.cloudera.org:8080/view/Bigtop-trunk/job/Bigtop-trunk-runsmokes/59/testReport/org.apache.bigtop.itest.hivesmoke/TestHiveSmokeBulk/testHiveBulk_union3_/
>         What's peculiar here is that nothing has failed *before* or
> *after* these particular
>         tests. Hence I don't think that the state of the cluster
> deployment is to blame.

Both errors seemed to be related to Hive unit tests failing since MiniHDFSCluster didn't come up:

Job Submission failed with exception 'org.apache.hadoop.ipc.RemoteException(File /user/jenkins/.staging/job_1371348647940_0308/job.split could only be replicated to 0 nodes instead of minReplication (=1).  There are 4 datanode(s) running and no node(s) are excluded in this operation.
>      4. All of the Mahout tests failed with the following:
>              http://bigtop01.cloudera.org:8080/view/Bigtop-trunk/job/Bigtop-trunk-runsmokes/59/testReport/org.apache.bigtop.itest.mahout.smoke/TestMahoutExamples/clusterControlDataWithCanopy/
>          This odd, and as I said -- all I know at this point is that
> the very same
>           tests running the very same Mahout pass with Hadoop 2.0.5-alpha.

Again, this is related to fact that we need to recompile Mahout against 2.1.0-beta - in particular this was due to the compatibility work done via MAPREDUCE-5156 (part of MAPREDUCE-5108).