Hive
  1. Hive
  2. HIVE-6356

Dependency injection in hbase storage handler is broken

    Details

    • Type: Bug Bug
    • Status: Resolved
    • Priority: Minor Minor
    • Resolution: Fixed
    • Affects Version/s: None
    • Fix Version/s: 0.13.0
    • Component/s: HBase Handler
    • Labels:
      None

      Description

      Dependent jars for hbase is not added to tmpjars, which is caused by the change of method signature(TableMapReduceUtil.addDependencyJars).

      1. HIVE-6356.1.patch.txt
        3 kB
        Navis
      2. HIVE-6356.2.patch.txt
        2 kB
        Navis
      3. HIVE-6356.3.patch.txt
        2 kB
        Navis
      4. HIVE-6356.4.patch.txt
        2 kB
        Xuefu Zhang
      5. HIVE-6356.4.patch.txt
        2 kB
        Xuefu Zhang
      6. HIVE-6356.addendum.00.patch
        3 kB
        Nick Dimiduk

        Issue Links

          Activity

          Hide
          Hive QA added a comment -

          Overall: +1 all checks pass

          Here are the results of testing the latest attachment:
          https://issues.apache.org/jira/secure/attachment/12626614/HIVE-6356.1.patch.txt

          SUCCESS: +1 4997 tests passed

          Test results: http://bigtop01.cloudera.org:8080/job/PreCommit-HIVE-Build/1165/testReport
          Console output: http://bigtop01.cloudera.org:8080/job/PreCommit-HIVE-Build/1165/console

          Messages:

          Executing org.apache.hive.ptest.execution.PrepPhase
          Executing org.apache.hive.ptest.execution.ExecutionPhase
          Executing org.apache.hive.ptest.execution.ReportingPhase
          

          This message is automatically generated.

          ATTACHMENT ID: 12626614

          Show
          Hive QA added a comment - Overall : +1 all checks pass Here are the results of testing the latest attachment: https://issues.apache.org/jira/secure/attachment/12626614/HIVE-6356.1.patch.txt SUCCESS: +1 4997 tests passed Test results: http://bigtop01.cloudera.org:8080/job/PreCommit-HIVE-Build/1165/testReport Console output: http://bigtop01.cloudera.org:8080/job/PreCommit-HIVE-Build/1165/console Messages: Executing org.apache.hive.ptest.execution.PrepPhase Executing org.apache.hive.ptest.execution.ExecutionPhase Executing org.apache.hive.ptest.execution.ReportingPhase This message is automatically generated. ATTACHMENT ID: 12626614
          Hide
          Ashutosh Chauhan added a comment -

          Is htrace strictly required ? If so, than don't we need to make sure its jar is available at run-time (currently it seems we don't have it in our lib/ dir of package?) Nick Dimiduk Can you also take a look?

          Show
          Ashutosh Chauhan added a comment - Is htrace strictly required ? If so, than don't we need to make sure its jar is available at run-time (currently it seems we don't have it in our lib/ dir of package?) Nick Dimiduk Can you also take a look?
          Hide
          Nick Dimiduk added a comment -

          I stumbled into this recently as well. HTrace is now a required runtime dependency, even when it's not used. This patch is incorrect, however. Because Hive is using the mapred namespace classes, the correct API is to invoke o.a.h.hbase.mapred.TableMapReduceUtil#addDependencyJars(JobConf). This will wire in all of HBase's runtime dependencies for you, and also attempt to auto-detect additional dependencies based on the JobConf (output classes, partitioners, formats, etc). If you want more fine-grained control over these dependencies (as Pig did, see PIG-3285), there are additional static methods in the o.a.h.hbase.mapreduce.TableMapReduceUtil class.

          For Hive's purpose, I think you'll be fine with just calling mapred.TableMapReduceUtil#addDependencyJars(JobConf). Having a smoke test that runs in pseudo-distributed mode would be helpful in verifying all requirements are met.

          Show
          Nick Dimiduk added a comment - I stumbled into this recently as well. HTrace is now a required runtime dependency, even when it's not used. This patch is incorrect, however. Because Hive is using the mapred namespace classes, the correct API is to invoke o.a.h.hbase.mapred.TableMapReduceUtil#addDependencyJars(JobConf). This will wire in all of HBase's runtime dependencies for you, and also attempt to auto-detect additional dependencies based on the JobConf (output classes, partitioners, formats, etc). If you want more fine-grained control over these dependencies (as Pig did, see PIG-3285 ), there are additional static methods in the o.a.h.hbase.mapreduce.TableMapReduceUtil class. For Hive's purpose, I think you'll be fine with just calling mapred.TableMapReduceUtil#addDependencyJars(JobConf). Having a smoke test that runs in pseudo-distributed mode would be helpful in verifying all requirements are met.
          Hide
          Navis added a comment -

          Right. I've forgot there are two version of TableMapReduceUtil. HIVE-3603 changed import clause of TableMapReduceUtil to mapred, which cause this problem. I'll fix this shortly after.

          Show
          Navis added a comment - Right. I've forgot there are two version of TableMapReduceUtil. HIVE-3603 changed import clause of TableMapReduceUtil to mapred, which cause this problem. I'll fix this shortly after.
          Hide
          Hive QA added a comment -

          Overall: +1 all checks pass

          Here are the results of testing the latest attachment:
          https://issues.apache.org/jira/secure/attachment/12626785/HIVE-6356.2.patch.txt

          SUCCESS: +1 4997 tests passed

          Test results: http://bigtop01.cloudera.org:8080/job/PreCommit-HIVE-Build/1172/testReport
          Console output: http://bigtop01.cloudera.org:8080/job/PreCommit-HIVE-Build/1172/console

          Messages:

          Executing org.apache.hive.ptest.execution.PrepPhase
          Executing org.apache.hive.ptest.execution.ExecutionPhase
          Executing org.apache.hive.ptest.execution.ReportingPhase
          

          This message is automatically generated.

          ATTACHMENT ID: 12626785

          Show
          Hive QA added a comment - Overall : +1 all checks pass Here are the results of testing the latest attachment: https://issues.apache.org/jira/secure/attachment/12626785/HIVE-6356.2.patch.txt SUCCESS: +1 4997 tests passed Test results: http://bigtop01.cloudera.org:8080/job/PreCommit-HIVE-Build/1172/testReport Console output: http://bigtop01.cloudera.org:8080/job/PreCommit-HIVE-Build/1172/console Messages: Executing org.apache.hive.ptest.execution.PrepPhase Executing org.apache.hive.ptest.execution.ExecutionPhase Executing org.apache.hive.ptest.execution.ReportingPhase This message is automatically generated. ATTACHMENT ID: 12626785
          Hide
          Nick Dimiduk added a comment -

          Patch 2 looks better. You should be able to do without explicitly requesting org.apache.hadoop.hbase.HBaseConfiguration.class as well; this should be included by the underlying HBase method.

          +1 (for what it's worth)

          Show
          Nick Dimiduk added a comment - Patch 2 looks better. You should be able to do without explicitly requesting org.apache.hadoop.hbase.HBaseConfiguration.class as well; this should be included by the underlying HBase method. +1 (for what it's worth)
          Hide
          Ashutosh Chauhan added a comment -

          +1

          Show
          Ashutosh Chauhan added a comment - +1
          Hide
          Ashutosh Chauhan added a comment -

          Committed to trunk. Thanks, Navis!

          Show
          Ashutosh Chauhan added a comment - Committed to trunk. Thanks, Navis!
          Hide
          Navis added a comment -

          Sadly, patch2 cannot resolve this issue. I should I have done some tests before updating the patch (the problem looked so obvious, sorry). o.a.h.hbase.mapred.TableMapReduceUtil#addDependencyJars(JobConf) is only works with older version of hbase.

          Now back to the problem, o.a.h.hbase.mapreduce.TableMapReduceUtil takes Job as parameter rather than JobConf, which seemed not good idea thinking that it's mere a wrapper of JobConf. Anyway, we can resolve this with patch1 (copy needed class list from TableMapReduceUtil) or make new Job instance and copy values from it to original JobConf. Which one is better?

          Show
          Navis added a comment - Sadly, patch2 cannot resolve this issue. I should I have done some tests before updating the patch (the problem looked so obvious, sorry). o.a.h.hbase.mapred.TableMapReduceUtil#addDependencyJars(JobConf) is only works with older version of hbase. Now back to the problem, o.a.h.hbase.mapreduce.TableMapReduceUtil takes Job as parameter rather than JobConf, which seemed not good idea thinking that it's mere a wrapper of JobConf. Anyway, we can resolve this with patch1 (copy needed class list from TableMapReduceUtil) or make new Job instance and copy values from it to original JobConf. Which one is better?
          Hide
          Nick Dimiduk added a comment -

          Please don't copy-paste the dependency list from HBase. The whole point of the addDependencyJars is to encapsulate that list in a single place that's managed by the HBase team.

          The reason the method accepts a Job instead of just a Configuration is that it inspects the job to add additional dependencies, such as input/output formats. Please refresh your memory from the comments on HIVE-2379.

          I'm attaching an addendum that should fix this. To be sure, I'll test it out on a real cluster later this afternoon. If the Hive team doesn't want the additional input/output format snooping, we can use an alternative method I introduced in HBASE-9165 for the benefit of PIG-3285.

          Show
          Nick Dimiduk added a comment - Please don't copy-paste the dependency list from HBase. The whole point of the addDependencyJars is to encapsulate that list in a single place that's managed by the HBase team. The reason the method accepts a Job instead of just a Configuration is that it inspects the job to add additional dependencies, such as input/output formats. Please refresh your memory from the comments on HIVE-2379 . I'm attaching an addendum that should fix this. To be sure, I'll test it out on a real cluster later this afternoon. If the Hive team doesn't want the additional input/output format snooping, we can use an alternative method I introduced in HBASE-9165 for the benefit of PIG-3285 .
          Hide
          Sergey Shelukhin added a comment -

          addendum looks good to me

          Show
          Sergey Shelukhin added a comment - addendum looks good to me
          Hide
          Ashutosh Chauhan added a comment -

          Navis Can you reupload the patch with correct name so Hive QA picks it up.

          Show
          Ashutosh Chauhan added a comment - Navis Can you reupload the patch with correct name so Hive QA picks it up.
          Hide
          Navis added a comment -

          addendum is basically same with already applied patch(patch2) and it's not fixing this problem. If copying is that bad, I should use second method to fix this.

          Show
          Navis added a comment - addendum is basically same with already applied patch(patch2) and it's not fixing this problem. If copying is that bad, I should use second method to fix this.
          Hide
          Nick Dimiduk added a comment -

          Patch 3 is not the addendum. My mistake, I thought patch2 had been committed.

          it's not fixing this problem

          What exactly is the problem it's not fixing? Is the job failing due to missing dependencies? Which dependencies? Are they on the classpath of the job client?

          I think copying really is that bad; it means this integration will break every time HBase changes it's dependencies. Not often, but it does happen. Instead, use the provided method and the correct dependencies will be added at runtime, everytime. Is this behavior undesirable for some reason?

          What do you mean "second method"?

          Show
          Nick Dimiduk added a comment - Patch 3 is not the addendum. My mistake, I thought patch2 had been committed. it's not fixing this problem What exactly is the problem it's not fixing? Is the job failing due to missing dependencies? Which dependencies? Are they on the classpath of the job client? I think copying really is that bad; it means this integration will break every time HBase changes it's dependencies. Not often, but it does happen. Instead, use the provided method and the correct dependencies will be added at runtime, everytime. Is this behavior undesirable for some reason? What do you mean "second method"?
          Hide
          Navis added a comment -

          Is the job failing due to missing dependencies

          Yes, it fails with ClassNotFoundException.

          org.apache.hadoop.hbase.mapred.TableMapReduceUtil.addDependencyJars(jobConf)
          org.apache.hadoop.hbase.mapreduce.TableMapReduceUtil.addDependencyJars(jobConf, HBaseStorageHandler.class)
          

          This adds zookeeper, netty, guava, hadoop, hive-exec, hive-hbase-handler to tmpJars, missing hbase related classes.

          I think copying really is that bad

          Agree. Just wanted quick, easy fix.

          What do you mean "second method"?

          In my comment above, "Anyway, we can resolve this with patch1 (copy needed class list from TableMapReduceUtil) or make new Job instance and copy values from it to original JobConf". You can see it in patch3.

          Show
          Navis added a comment - Is the job failing due to missing dependencies Yes, it fails with ClassNotFoundException. org.apache.hadoop.hbase.mapred.TableMapReduceUtil.addDependencyJars(jobConf) org.apache.hadoop.hbase.mapreduce.TableMapReduceUtil.addDependencyJars(jobConf, HBaseStorageHandler.class) This adds zookeeper, netty, guava, hadoop, hive-exec, hive-hbase-handler to tmpJars, missing hbase related classes. I think copying really is that bad Agree. Just wanted quick, easy fix. What do you mean "second method"? In my comment above, "Anyway, we can resolve this with patch1 (copy needed class list from TableMapReduceUtil) or make new Job instance and copy values from it to original JobConf". You can see it in patch3.
          Hide
          Nick Dimiduk added a comment -

          Which version of HBase are you testing against? mapred.TableMapReduceUtil#addDependencyJars(JobConf) calls the same method from the mapreduce package, which in turn adds all the hbase jars, plus the hbase dependencies you mention.

          ClassNotFoundException thrown at the job client or in the job? At the client means your local classpath is missing something. In the job means it wasn't packaged properly. What's the missing class/jar?

          Show
          Nick Dimiduk added a comment - Which version of HBase are you testing against? mapred.TableMapReduceUtil#addDependencyJars(JobConf) calls the same method from the mapreduce package, which in turn adds all the hbase jars, plus the hbase dependencies you mention. ClassNotFoundException thrown at the job client or in the job? At the client means your local classpath is missing something. In the job means it wasn't packaged properly. What's the missing class/jar?
          Hide
          Navis added a comment -

          Ah, you are using hbase-0.96.1, which is applied HBASE-9165. I got it. Hive in trunk uses hbase-0.96.0 (addHBaseDependencyJars() does not exist in it).

          Should we upgrade hbase to 0.96.1? I prefer to support both versions by applying patch3 but I'm good with any way.

          Show
          Navis added a comment - Ah, you are using hbase-0.96.1, which is applied HBASE-9165 . I got it. Hive in trunk uses hbase-0.96.0 (addHBaseDependencyJars() does not exist in it). Should we upgrade hbase to 0.96.1? I prefer to support both versions by applying patch3 but I'm good with any way.
          Hide
          Nick Dimiduk added a comment -

          With my HBase hat on, I'd say the upgrade to 0.96.1 us prudent for both the API improvements and the performance improvements. Not sure how Hive folks feel about this.

          Show
          Nick Dimiduk added a comment - With my HBase hat on, I'd say the upgrade to 0.96.1 us prudent for both the API improvements and the performance improvements. Not sure how Hive folks feel about this.
          Hide
          Nick Dimiduk added a comment -

          Even on 0.96.0, both mapred.TableMapReduceUtil#addDependencyJars(JobConf) and mapreduce.TableMapReduceUtil#addDependencyJars(Job) handle the hbase dependencies for you. I'm back to not understanding what issue you're having

          Show
          Nick Dimiduk added a comment - Even on 0.96.0, both mapred.TableMapReduceUtil#addDependencyJars(JobConf) and mapreduce.TableMapReduceUtil#addDependencyJars(Job) handle the hbase dependencies for you. I'm back to not understanding what issue you're having
          Hide
          Ashutosh Chauhan added a comment -

          I am also in favor of bumping HBase version. We already have code complexity supporting multiple versions of Hadoop, we don't worsen that by supporting multiple HBase version.

          Show
          Ashutosh Chauhan added a comment - I am also in favor of bumping HBase version. We already have code complexity supporting multiple versions of Hadoop, we don't worsen that by supporting multiple HBase version.
          Hide
          Navis added a comment -

          Even on 0.96.0, both mapred.TableMapReduceUtil#addDependencyJars(JobConf) and mapreduce.TableMapReduceUtil#addDependencyJars(Job) handle the hbase dependencies for you

          Really wish to call the method but all we got is JobConf, not Job and there is no way just to wrap existing JobConf(copies in it) with Job. So in patch3, I've create dummy Job instance to call it and copied values(in "tmpJars") in it to original JobConf. I've mentioned this once already and even quoted again as you requested explanation.

          o.a.h.hbase.mapreduce.TableMapReduceUtil takes Job as parameter rather than JobConf

          "..or make new Job instance and copy values from it to original JobConf". You can see it in patch3.

          I'm ok with fixing this in anyway (upgrade hbase to 0.96.1 or applying patch3), though I prefer the later way to support hbase-0.96.0, which is still in use by many companies.

          Show
          Navis added a comment - Even on 0.96.0, both mapred.TableMapReduceUtil#addDependencyJars(JobConf) and mapreduce.TableMapReduceUtil#addDependencyJars(Job) handle the hbase dependencies for you Really wish to call the method but all we got is JobConf, not Job and there is no way just to wrap existing JobConf(copies in it) with Job. So in patch3, I've create dummy Job instance to call it and copied values(in "tmpJars") in it to original JobConf. I've mentioned this once already and even quoted again as you requested explanation. o.a.h.hbase.mapreduce.TableMapReduceUtil takes Job as parameter rather than JobConf "..or make new Job instance and copy values from it to original JobConf". You can see it in patch3. I'm ok with fixing this in anyway (upgrade hbase to 0.96.1 or applying patch3), though I prefer the later way to support hbase-0.96.0, which is still in use by many companies.
          Hide
          Nick Dimiduk added a comment -

          all we got is JobConf, not Job

          So why not call mapred.TableMapReduceUtil#addDependencyJars(JobConf) ? It does what you need; that's why it's there.

          Show
          Nick Dimiduk added a comment - all we got is JobConf, not Job So why not call mapred.TableMapReduceUtil#addDependencyJars(JobConf) ? It does what you need; that's why it's there.
          Hide
          Navis added a comment -

          In hbase-0.96.0, mapred.TableMapReduceUtil#addDependencyJars(JobConf) does not call "org.apache.hadoop.hbase.mapreduce.TableMapReduceUtil.addHBaseDependencyJars(job)" which is not exist.

          public static void addDependencyJars(JobConf job) throws IOException {
              org.apache.hadoop.hbase.mapreduce.TableMapReduceUtil.addDependencyJars(
                job,
                org.apache.zookeeper.ZooKeeper.class,
                org.jboss.netty.channel.ChannelFactory.class,
                com.google.common.base.Function.class,
                com.google.protobuf.Message.class,
                job.getMapOutputKeyClass(),
                job.getMapOutputValueClass(),
                job.getOutputKeyClass(),
                job.getOutputValueClass(),
                job.getPartitionerClass(),
                job.getClass("mapred.input.format.class", TextInputFormat.class, InputFormat.class),
                job.getClass("mapred.output.format.class", TextOutputFormat.class, OutputFormat.class),
                job.getCombinerClass());
            }
          

          Again, this only

          adds zookeeper, netty, guava, hadoop, hive-exec, hive-hbase-handler to tmpJars, missing hbase related classes.

          Show
          Navis added a comment - In hbase-0.96.0, mapred.TableMapReduceUtil#addDependencyJars(JobConf) does not call "org.apache.hadoop.hbase.mapreduce.TableMapReduceUtil.addHBaseDependencyJars(job)" which is not exist. public static void addDependencyJars(JobConf job) throws IOException { org.apache.hadoop.hbase.mapreduce.TableMapReduceUtil.addDependencyJars( job, org.apache.zookeeper.ZooKeeper.class, org.jboss.netty.channel.ChannelFactory.class, com.google.common.base.Function.class, com.google.protobuf.Message.class, job.getMapOutputKeyClass(), job.getMapOutputValueClass(), job.getOutputKeyClass(), job.getOutputValueClass(), job.getPartitionerClass(), job.getClass( "mapred.input.format.class" , TextInputFormat.class, InputFormat.class), job.getClass( "mapred.output.format.class" , TextOutputFormat.class, OutputFormat.class), job.getCombinerClass()); } Again, this only adds zookeeper, netty, guava, hadoop, hive-exec, hive-hbase-handler to tmpJars, missing hbase related classes.
          Hide
          Nick Dimiduk added a comment -

          Yes, of course; I now see the issue. My sincerest apologies for belaboring an invalid point

          In the interest of supporting as many users as possible, patch v3 seems the most prudent approach, with a note about deferring back to HBase methods once Hive moves on to 0.98.0.

          Show
          Nick Dimiduk added a comment - Yes, of course; I now see the issue. My sincerest apologies for belaboring an invalid point In the interest of supporting as many users as possible, patch v3 seems the most prudent approach, with a note about deferring back to HBase methods once Hive moves on to 0.98.0.
          Hide
          Ashutosh Chauhan added a comment -

          Marking this as Patch Available. Lets get this in. Code changes are rather small to bump up hbase version.

          Show
          Ashutosh Chauhan added a comment - Marking this as Patch Available. Lets get this in. Code changes are rather small to bump up hbase version.
          Hide
          Xuefu Zhang added a comment -

          Ashutosh Chauhan Could we put a close on this? I understood that patch v3 is committed to trunk. Do we still need addendum patch to be committed, in order to close this JIRA?

          Show
          Xuefu Zhang added a comment - Ashutosh Chauhan Could we put a close on this? I understood that patch v3 is committed to trunk. Do we still need addendum patch to be committed, in order to close this JIRA?
          Hide
          Xuefu Zhang added a comment -

          If we all agree that patch v3 is better, should we revert the following commit (which is patch v2) and commit patch v3 instead?

          commit 57467b4fd8cc5acb5250b635f86c3115691cf367
          Author: Ashutosh Chauhan <hashutosh@apache.org>
          Date: Tue Feb 4 23:42:30 2014 +0000

          HIVE-6356 : Dependency injection in hbase storage handler is broken (Navis v

          Show
          Xuefu Zhang added a comment - If we all agree that patch v3 is better, should we revert the following commit (which is patch v2) and commit patch v3 instead? commit 57467b4fd8cc5acb5250b635f86c3115691cf367 Author: Ashutosh Chauhan <hashutosh@apache.org> Date: Tue Feb 4 23:42:30 2014 +0000 HIVE-6356 : Dependency injection in hbase storage handler is broken (Navis v
          Hide
          Ashutosh Chauhan added a comment -

          I think v3 is required in addition to v2. I am fine with committing v3 (with a comment in the code that problem is fixed in 0.96.1 and later, so those changes will not be required once Hive bumps up its hbase version).

          Show
          Ashutosh Chauhan added a comment - I think v3 is required in addition to v2. I am fine with committing v3 (with a comment in the code that problem is fixed in 0.96.1 and later, so those changes will not be required once Hive bumps up its hbase version).
          Hide
          Xuefu Zhang added a comment -

          Thanks for the clarification, Ashutosh Chauhan. I will update the patch v3 with the necessary comments.

          Show
          Xuefu Zhang added a comment - Thanks for the clarification, Ashutosh Chauhan . I will update the patch v3 with the necessary comments.
          Hide
          Xuefu Zhang added a comment -

          Patch #4 is basically #3 with necessary comments.

          Show
          Xuefu Zhang added a comment - Patch #4 is basically #3 with necessary comments.
          Hide
          Navis added a comment -

          +1

          Show
          Navis added a comment - +1
          Hide
          Hive QA added a comment -

          Overall: -1 at least one tests failed

          Here are the results of testing the latest attachment:
          https://issues.apache.org/jira/secure/attachment/12631340/HIVE-6356.4.patch.txt

          ERROR: -1 due to 1 failed/errored test(s), 5180 tests executed
          Failed tests:

          org.apache.hcatalog.hbase.snapshot.lock.TestWriteLock.testRun
          

          Test results: http://bigtop01.cloudera.org:8080/job/PreCommit-HIVE-Build/1551/testReport
          Console output: http://bigtop01.cloudera.org:8080/job/PreCommit-HIVE-Build/1551/console

          Messages:

          Executing org.apache.hive.ptest.execution.PrepPhase
          Executing org.apache.hive.ptest.execution.ExecutionPhase
          Executing org.apache.hive.ptest.execution.ReportingPhase
          Tests exited with: TestsFailedException: 1 tests failed
          

          This message is automatically generated.

          ATTACHMENT ID: 12631340

          Show
          Hive QA added a comment - Overall : -1 at least one tests failed Here are the results of testing the latest attachment: https://issues.apache.org/jira/secure/attachment/12631340/HIVE-6356.4.patch.txt ERROR: -1 due to 1 failed/errored test(s), 5180 tests executed Failed tests: org.apache.hcatalog.hbase.snapshot.lock.TestWriteLock.testRun Test results: http://bigtop01.cloudera.org:8080/job/PreCommit-HIVE-Build/1551/testReport Console output: http://bigtop01.cloudera.org:8080/job/PreCommit-HIVE-Build/1551/console Messages: Executing org.apache.hive.ptest.execution.PrepPhase Executing org.apache.hive.ptest.execution.ExecutionPhase Executing org.apache.hive.ptest.execution.ReportingPhase Tests exited with: TestsFailedException: 1 tests failed This message is automatically generated. ATTACHMENT ID: 12631340
          Hide
          Xuefu Zhang added a comment -

          The above test seems transient. I manually re-ran it and it passed. Reattach the same patch (v4) for another run.

          Show
          Xuefu Zhang added a comment - The above test seems transient. I manually re-ran it and it passed. Reattach the same patch (v4) for another run.
          Hide
          Hive QA added a comment -

          Overall: +1 all checks pass

          Here are the results of testing the latest attachment:
          https://issues.apache.org/jira/secure/attachment/12631814/HIVE-6356.4.patch.txt

          SUCCESS: +1 5185 tests passed

          Test results: http://bigtop01.cloudera.org:8080/job/PreCommit-HIVE-Build/1566/testReport
          Console output: http://bigtop01.cloudera.org:8080/job/PreCommit-HIVE-Build/1566/console

          Messages:

          Executing org.apache.hive.ptest.execution.PrepPhase
          Executing org.apache.hive.ptest.execution.ExecutionPhase
          Executing org.apache.hive.ptest.execution.ReportingPhase
          

          This message is automatically generated.

          ATTACHMENT ID: 12631814

          Show
          Hive QA added a comment - Overall : +1 all checks pass Here are the results of testing the latest attachment: https://issues.apache.org/jira/secure/attachment/12631814/HIVE-6356.4.patch.txt SUCCESS: +1 5185 tests passed Test results: http://bigtop01.cloudera.org:8080/job/PreCommit-HIVE-Build/1566/testReport Console output: http://bigtop01.cloudera.org:8080/job/PreCommit-HIVE-Build/1566/console Messages: Executing org.apache.hive.ptest.execution.PrepPhase Executing org.apache.hive.ptest.execution.ExecutionPhase Executing org.apache.hive.ptest.execution.ReportingPhase This message is automatically generated. ATTACHMENT ID: 12631814
          Hide
          Xuefu Zhang added a comment -

          Patch committed to trunk. Thanks to Navis, Ashutosh, et al.

          Show
          Xuefu Zhang added a comment - Patch committed to trunk. Thanks to Navis, Ashutosh, et al.

            People

            • Assignee:
              Xuefu Zhang
              Reporter:
              Navis
            • Votes:
              0 Vote for this issue
              Watchers:
              4 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved:

                Development