Re: Failing tests, precommit test runtimes

[Siddharth Seth]

Created HIVE-13402 to disable the tests. Have narrowed it down to 3 hanging
tests. From what I understand the TestVectorTimestampExpressions tests have
already been fixed.

On Tue, Mar 29, 2016 at 10:27 AM, Ashutosh Chauhan 
wrote:

> I agree with disabling these tests till we are able to fix it. Since bug
> seems to be in spark runtime, it may take a little while to fix it.
> Increased run time for QA till then would be waste of resources.
>
> Thanks,
> Ashutosh
>
> On Tue, Mar 29, 2016 at 9:06 AM, Siddharth Seth  wrote:
>
> > Thanks Szehon. Should we disable the tests while the issue is resolved.
> > There's a backlog of 17 patches - which will take over 2 days to run at
> > this point.
> > Also, HIVE-13365 ended up running for over 8 hours yesterday. Maybe we
> > should put a cap on max runtime for a single run. Anyone know how to do
> > this ?
> >
> > On Mon, Mar 28, 2016 at 3:05 PM, Szehon Ho  wrote:
> >
> > > Yea we looked at it before.. its tracked by HIVE-13223, caused by
> > > HIVE-13040.
> > >
> > > Because it hangs these tests take 2 hours max now before they are
> killed,
> > > hence each run taking over 3 hours.
> > >
> > > Ashutosh is looking I believe, I havent had a chance to take a look
> > myself.
> > >
> > > Szehon
> > >
> > > On Mon, Mar 28, 2016 at 2:28 PM, Siddharth Seth 
> > wrote:
> > >
> > > > There's several tests which are failing consistently. Should we
> disable
> > > > these and create blocker jiras to fix and enable thees for the next
> > > > release. The list of test is at the end of the mail.
> > > >
> > > > Also, each pre-commit run seems to be taking over 3 hours. Is this
> > > normal ?
> > > > Having test results come back 2 days after submitting a patch is not
> > very
> > > > efficient.
> > > >
> > > > Thanks,
> > > > Sid
> > > >
> > > > TestSparkCliDriver-groupby3_map.q-sample2.q-auto_join14.q-and-12-more
> > > > - did not produce a TEST-*.xml file
> > > >
> > > >
> > >
> >
> TestSparkCliDriver-groupby_map_ppr_multi_distinct.q-table_access_keys_stats.q-groupby4_noskew.q-and-12-more
> > > > - did not produce a TEST-*.xml file
> > > >
> > > >
> > >
> >
> TestSparkCliDriver-join_rc.q-insert1.q-vectorized_rcfile_columnar.q-and-12-more
> > > > - did not produce a TEST-*.xml file
> > > > TestSparkCliDriver-ppd_join4.q-join9.q-ppd_join3.q-and-12-more - did
> > > > not produce a TEST-*.xml file
> > > >
> > > >
> > >
> >
> TestSparkCliDriver-timestamp_lazy.q-bucketsortoptimize_insert_4.q-date_udf.q-and-12-more
> > > > - did not produce a TEST-*.xml file
> > > >
> > > >
> > > >
> > >
> >
> org.apache.hadoop.hive.ql.exec.vector.expressions.TestVectorTimestampExpressions.testVectorUDFMonthString
> > > >
> > > >
> > >
> >
> org.apache.hadoop.hive.ql.exec.vector.expressions.TestVectorTimestampExpressions.testVectorUDFMonthTimestamp
> > > >
> > > >
> > >
> >
> org.apache.hadoop.hive.ql.exec.vector.expressions.TestVectorTimestampExpressions.testVectorUDFYearString
> > > >
> > > >
> > >
> >
> org.apache.hadoop.hive.ql.exec.vector.expressions.TestVectorTimestampExpressions.testVectorUDFYearTimestamp
> > > >
> > >
> >
>

Re: Failing tests, precommit test runtimes

[Ashutosh Chauhan]

I agree with disabling these tests till we are able to fix it. Since bug
seems to be in spark runtime, it may take a little while to fix it.
Increased run time for QA till then would be waste of resources.

Thanks,
Ashutosh

On Tue, Mar 29, 2016 at 9:06 AM, Siddharth Seth  wrote:

> Thanks Szehon. Should we disable the tests while the issue is resolved.
> There's a backlog of 17 patches - which will take over 2 days to run at
> this point.
> Also, HIVE-13365 ended up running for over 8 hours yesterday. Maybe we
> should put a cap on max runtime for a single run. Anyone know how to do
> this ?
>
> On Mon, Mar 28, 2016 at 3:05 PM, Szehon Ho  wrote:
>
> > Yea we looked at it before.. its tracked by HIVE-13223, caused by
> > HIVE-13040.
> >
> > Because it hangs these tests take 2 hours max now before they are killed,
> > hence each run taking over 3 hours.
> >
> > Ashutosh is looking I believe, I havent had a chance to take a look
> myself.
> >
> > Szehon
> >
> > On Mon, Mar 28, 2016 at 2:28 PM, Siddharth Seth 
> wrote:
> >
> > > There's several tests which are failing consistently. Should we disable
> > > these and create blocker jiras to fix and enable thees for the next
> > > release. The list of test is at the end of the mail.
> > >
> > > Also, each pre-commit run seems to be taking over 3 hours. Is this
> > normal ?
> > > Having test results come back 2 days after submitting a patch is not
> very
> > > efficient.
> > >
> > > Thanks,
> > > Sid
> > >
> > > TestSparkCliDriver-groupby3_map.q-sample2.q-auto_join14.q-and-12-more
> > > - did not produce a TEST-*.xml file
> > >
> > >
> >
> TestSparkCliDriver-groupby_map_ppr_multi_distinct.q-table_access_keys_stats.q-groupby4_noskew.q-and-12-more
> > > - did not produce a TEST-*.xml file
> > >
> > >
> >
> TestSparkCliDriver-join_rc.q-insert1.q-vectorized_rcfile_columnar.q-and-12-more
> > > - did not produce a TEST-*.xml file
> > > TestSparkCliDriver-ppd_join4.q-join9.q-ppd_join3.q-and-12-more - did
> > > not produce a TEST-*.xml file
> > >
> > >
> >
> TestSparkCliDriver-timestamp_lazy.q-bucketsortoptimize_insert_4.q-date_udf.q-and-12-more
> > > - did not produce a TEST-*.xml file
> > >
> > >
> > >
> >
> org.apache.hadoop.hive.ql.exec.vector.expressions.TestVectorTimestampExpressions.testVectorUDFMonthString
> > >
> > >
> >
> org.apache.hadoop.hive.ql.exec.vector.expressions.TestVectorTimestampExpressions.testVectorUDFMonthTimestamp
> > >
> > >
> >
> org.apache.hadoop.hive.ql.exec.vector.expressions.TestVectorTimestampExpressions.testVectorUDFYearString
> > >
> > >
> >
> org.apache.hadoop.hive.ql.exec.vector.expressions.TestVectorTimestampExpressions.testVectorUDFYearTimestamp
> > >
> >
>

Re: Failing tests, precommit test runtimes

[Siddharth Seth]

Thanks Szehon. Should we disable the tests while the issue is resolved.
There's a backlog of 17 patches - which will take over 2 days to run at
this point.
Also, HIVE-13365 ended up running for over 8 hours yesterday. Maybe we
should put a cap on max runtime for a single run. Anyone know how to do
this ?

On Mon, Mar 28, 2016 at 3:05 PM, Szehon Ho  wrote:

> Yea we looked at it before.. its tracked by HIVE-13223, caused by
> HIVE-13040.
>
> Because it hangs these tests take 2 hours max now before they are killed,
> hence each run taking over 3 hours.
>
> Ashutosh is looking I believe, I havent had a chance to take a look myself.
>
> Szehon
>
> On Mon, Mar 28, 2016 at 2:28 PM, Siddharth Seth  wrote:
>
> > There's several tests which are failing consistently. Should we disable
> > these and create blocker jiras to fix and enable thees for the next
> > release. The list of test is at the end of the mail.
> >
> > Also, each pre-commit run seems to be taking over 3 hours. Is this
> normal ?
> > Having test results come back 2 days after submitting a patch is not very
> > efficient.
> >
> > Thanks,
> > Sid
> >
> > TestSparkCliDriver-groupby3_map.q-sample2.q-auto_join14.q-and-12-more
> > - did not produce a TEST-*.xml file
> >
> >
> TestSparkCliDriver-groupby_map_ppr_multi_distinct.q-table_access_keys_stats.q-groupby4_noskew.q-and-12-more
> > - did not produce a TEST-*.xml file
> >
> >
> TestSparkCliDriver-join_rc.q-insert1.q-vectorized_rcfile_columnar.q-and-12-more
> > - did not produce a TEST-*.xml file
> > TestSparkCliDriver-ppd_join4.q-join9.q-ppd_join3.q-and-12-more - did
> > not produce a TEST-*.xml file
> >
> >
> TestSparkCliDriver-timestamp_lazy.q-bucketsortoptimize_insert_4.q-date_udf.q-and-12-more
> > - did not produce a TEST-*.xml file
> >
> >
> >
> org.apache.hadoop.hive.ql.exec.vector.expressions.TestVectorTimestampExpressions.testVectorUDFMonthString
> >
> >
> org.apache.hadoop.hive.ql.exec.vector.expressions.TestVectorTimestampExpressions.testVectorUDFMonthTimestamp
> >
> >
> org.apache.hadoop.hive.ql.exec.vector.expressions.TestVectorTimestampExpressions.testVectorUDFYearString
> >
> >
> org.apache.hadoop.hive.ql.exec.vector.expressions.TestVectorTimestampExpressions.testVectorUDFYearTimestamp
> >
>

Re: Failing tests, precommit test runtimes

[Szehon Ho]

Yea we looked at it before.. its tracked by HIVE-13223, caused by
HIVE-13040.

Because it hangs these tests take 2 hours max now before they are killed,
hence each run taking over 3 hours.

Ashutosh is looking I believe, I havent had a chance to take a look myself.

Szehon

On Mon, Mar 28, 2016 at 2:28 PM, Siddharth Seth  wrote:

> There's several tests which are failing consistently. Should we disable
> these and create blocker jiras to fix and enable thees for the next
> release. The list of test is at the end of the mail.
>
> Also, each pre-commit run seems to be taking over 3 hours. Is this normal ?
> Having test results come back 2 days after submitting a patch is not very
> efficient.
>
> Thanks,
> Sid
>
> TestSparkCliDriver-groupby3_map.q-sample2.q-auto_join14.q-and-12-more
> - did not produce a TEST-*.xml file
>
> TestSparkCliDriver-groupby_map_ppr_multi_distinct.q-table_access_keys_stats.q-groupby4_noskew.q-and-12-more
> - did not produce a TEST-*.xml file
>
> TestSparkCliDriver-join_rc.q-insert1.q-vectorized_rcfile_columnar.q-and-12-more
> - did not produce a TEST-*.xml file
> TestSparkCliDriver-ppd_join4.q-join9.q-ppd_join3.q-and-12-more - did
> not produce a TEST-*.xml file
>
> TestSparkCliDriver-timestamp_lazy.q-bucketsortoptimize_insert_4.q-date_udf.q-and-12-more
> - did not produce a TEST-*.xml file
>
>
> org.apache.hadoop.hive.ql.exec.vector.expressions.TestVectorTimestampExpressions.testVectorUDFMonthString
>
> org.apache.hadoop.hive.ql.exec.vector.expressions.TestVectorTimestampExpressions.testVectorUDFMonthTimestamp
>
> org.apache.hadoop.hive.ql.exec.vector.expressions.TestVectorTimestampExpressions.testVectorUDFYearString
>
> org.apache.hadoop.hive.ql.exec.vector.expressions.TestVectorTimestampExpressions.testVectorUDFYearTimestamp
>