I agree that there should be tests with queuing enabled (at least sanity 
tests). I did not mean to delay the release, but this regression causes all 
queries to fail with an illegal state transition exception (when queueing is 
enabled).

Thank you,
Sudheesh

> On Feb 8, 2016, at 6:22 PM, Jason Altekruse <[email protected]> wrote:
> 
> The case that was reported in the JIRA was a failure on a very simple
> query:  select * from sys.options;
> 
> I assume this means that any query will fail when queuing is enabled. That
> would make a strong case for inclusion in the release, I didn't look
> closely at the JIRA before. Hakim, you reviewed the patch, but it doesn't
> include any new tests. Did Hanifi mention if the change made there was
> necessary to pretty much fix any query when queuing was enabled?
> 
> - Jason
> 
> On Mon, Feb 8, 2016 at 4:57 PM, Abdel Hakim Deneche <[email protected]>
> wrote:
> 
>> Does it mean that any user who's been using queuing won't be able to use
>> 1.5.0 ?
>> 
>> On Mon, Feb 8, 2016 at 4:40 PM, Jason Altekruse <[email protected]>
>> wrote:
>> 
>>> Hey Sudheesh,
>>> 
>>> I just pushed Venki's fix for the Web UI issue to the master branch.
>>> 
>>> My fix for the build issue I ran into when trying to prepare the release
>> is
>>> a fair point. The change only has a very limited impact on the build, and
>>> only changes the result when running a release itself. I should have been
>>> better communicating the change that was made, I have the posted an
>> update
>>> on the JIRA I filed to do a follow-up investigation of the problem [1]. I
>>> didn't include it on m merge branch with Venki's change, but I will post
>> it
>>> shortly associated with this new JIRA [2] for review and kick off the
>> tests
>>> with the change rebased.
>>> 
>>> As far as 4235 is concerned, I would like the release to be as stable as
>>> possible, but the release has taken quite a long time to get to vote.
>> This
>>> issue was filed at the end of December, and was fixed just 4 days ago,
>> with
>>> no comment on the previous release thread about including the fix in the
>>> release. I fully support making queuing a first-class feature of Drill,
>> but
>>> we need to add automated tests for it if we want it to stay stable.
>>> 
>>> I'm open to discussion on the topic, but I'm not sure we should delay the
>>> release further for it.
>>> 
>>> - Jason
>>> 
>>> [1] - https://issues.apache.org/jira/browse/DRILL-4336
>>> [2] - https://issues.apache.org/jira/browse/DRILL-4375
>>> 
>>> On Mon, Feb 8, 2016 at 2:59 PM, Sudheesh Katkam <[email protected]>
>>> wrote:
>>> 
>>>> Although my vote is non-binding <
>>>> http://drill.apache.org/docs/project-bylaws/#actions>, I have two
>>>> concerns:
>>>> 
>>>> * DRILL-4187 <https://issues.apache.org/jira/browse/DRILL-4187>
>> caused a
>>>> critical regression noted in DRILL-4235 <
>>>> https://issues.apache.org/jira/browse/DRILL-4235>. There is a patch
>> for
>>>> DRILL-4235, which is not part of the release candidate. This can cause
>>>> failures for users that are using the queuing feature.
>>>> 
>>>> * There are commits made to the release branch <
>>>> https://github.com/jaltekruse/incubator-drill/commits/1.5-release-rc2>
>>> in
>>>> Jason's repo that are not checked in to master.
>>>> 
>>>> Thanks,
>>>> Sudheesh
>>>> 
>>>>> On Feb 8, 2016, at 2:30 PM, Jason Altekruse <
>> [email protected]>
>>>> wrote:
>>>>> 
>>>>> Thanks everyone who has voted so far. The vote closes tomorrow
>> morning
>>>> and
>>>>> right now we're only at the minimum number of binding votes for it to
>>>> pass.
>>>>> Anyone who has some time available, please try out the release and
>>> cast a
>>>>> vote.
>>>>> 
>>>>> On Mon, Feb 8, 2016 at 2:02 PM, Jacques Nadeau <[email protected]>
>>>> wrote:
>>>>> 
>>>>>> Downloaded, built and ran unit tests.
>>>>>> Manually tried a few queries.
>>>>>> 
>>>>>> Looks good
>>>>>> 
>>>>>> +1 (binding)
>>>>>> 
>>>>>> 
>>>>>> --
>>>>>> Jacques Nadeau
>>>>>> CTO and Co-Founder, Dremio
>>>>>> 
>>>>>> On Sun, Feb 7, 2016 at 10:03 AM, Aman Sinha <[email protected]>
>>>> wrote:
>>>>>> 
>>>>>>> +1
>>>>>>> - Downloaded src and built, ran unit tests on my Mac
>>>>>>> - Manually ran a few queries against TPC-DS
>>>>>>> - Verified partition pruning, metadata caching was working as
>>> expected
>>>>>> for
>>>>>>> these test queries
>>>>>>> - Checked query profile in Web UI, checked query cancellation
>>>>>>> - Found 1 performance issue with lots of small parquet files
>> ...filed
>>>>>>> DRILL-4365 but need confirmation whether it is reproducible for
>> other
>>>>>>> folks.  At this point, I am not considering it a blocker due to the
>>>> fact
>>>>>> I
>>>>>>> could not reproduce with a more general/bigger dataset.
>>>>>>> 
>>>>>>> Aman
>>>>>>> 
>>>>>>> On Fri, Feb 5, 2016 at 12:21 PM, Julien Le Dem <[email protected]>
>>>>>> wrote:
>>>>>>> 
>>>>>>>> +1 (non-binding)
>>>>>>>> Built and run the tests on linux (took 27 min)
>>>>>>>> 
>>>>>>>> 
>>>>>>>> 
>>>>>>>> On Fri, Feb 5, 2016 at 11:21 AM, Stefán Baxter <
>>>>>>> [email protected]>
>>>>>>>> wrote:
>>>>>>>> 
>>>>>>>>> +1 (non-binding / not a committer)
>>>>>>>>> 
>>>>>>>>>  - Built the project on ubuntu/linux
>>>>>>>>>  - Ran our test suite
>>>>>>>>>  - Verified that the jdbc driver works and is properly shaded
>> (we
>>>>>> had
>>>>>>>>>  problems with *leakage*)
>>>>>>>>> 
>>>>>>>>> (I ran into a problem reading a snappy zipped parquet file that
>> was
>>>>>>>> created
>>>>>>>>> with the latest parquet-mr/parquet-avro (1.8.1) but i think that
>> is
>>>>>> out
>>>>>>>> of
>>>>>>>>> scope here and I will create a Jira issue once I have tested it
>>>>>> better)
>>>>>>>>> 
>>>>>>>>> Thank you
>>>>>>>>> 
>>>>>>>>> On Fri, Feb 5, 2016 at 6:56 PM, Jason Altekruse <
>>>>>>>> [email protected]>
>>>>>>>>> wrote:
>>>>>>>>> 
>>>>>>>>>> Hello all,
>>>>>>>>>> 
>>>>>>>>>> I'd like to propose the third release candidate (rc2) of Apache
>>>>>>> Drill,
>>>>>>>>>> version
>>>>>>>>>> 1.5.0. It covers a total of 55 resolved JIRAs [1]. Thanks to
>>>>>> everyone
>>>>>>>> who
>>>>>>>>>> contributed to this release. This release candidate includes a
>> fix
>>>>>>> for
>>>>>>>>>> DRILL-4353, a major stability problem with the Rest API that was
>>>>>>>>> identified
>>>>>>>>>> during the last vote.
>>>>>>>>>> 
>>>>>>>>>> The tarball artifacts are hosted at [2] and the maven artifacts
>>> are
>>>>>>>>> hosted
>>>>>>>>>> at
>>>>>>>>>> [3]. This release candidate is based on commit
>>>>>>>>>> 0a64888ba8d374e94435e2518e81352e677255ad located at [4].
>>>>>>>>>> 
>>>>>>>>>> The vote will be open for the next 96 hours (including an extra
>>> day
>>>>>>> as
>>>>>>>>> the
>>>>>>>>>> vote is happening over a weekend) ending at 11AM Pacific,
>> February
>>>>>>> 9th,
>>>>>>>>>> 2016.
>>>>>>>>>> 
>>>>>>>>>> [ ] +1
>>>>>>>>>> [ ] +0
>>>>>>>>>> [ ] -1
>>>>>>>>>> 
>>>>>>>>>> Here's my vote: +1
>>>>>>>>>> 
>>>>>>>>>> Thanks,
>>>>>>>>>> Jason
>>>>>>>>>> 
>>>>>>>>>> [1]
>>>>>>>>>> 
>>>>>>>>>> 
>>>>>>>>> 
>>>>>>>> 
>>>>>>> 
>>>>>> 
>>>> 
>>> 
>> https://issues.apache.org/jira/secure/ReleaseNote.jspa?projectId=12313820&version=12332948
>>>>>>>>>> [2] http://people.apache.org/~json/apache-drill-1.5.0.rc2/
>>>>>>>>>> [3]
>>>>>>>>> 
>>>>>> 
>>> https://repository.apache.org/content/repositories/orgapachedrill-1026
>>>>>>>>>> [4]
>>>>>>> https://github.com/jaltekruse/incubator-drill/tree/1.5-release-rc2
>>>>>>>>>> 
>>>>>>>>> 
>>>>>>>> 
>>>>>>>> 
>>>>>>>> 
>>>>>>>> --
>>>>>>>> Julien
>>>>>>>> 
>>>>>>> 
>>>>>> 
>>>> 
>>>> 
>>> 
>> 
>> 
>> 
>> --
>> 
>> Abdelhakim Deneche
>> 
>> Software Engineer
>> 
>>  <http://www.mapr.com/>
>> 
>> 
>> Now Available - Free Hadoop On-Demand Training
>> <
>> http://www.mapr.com/training?utm_source=Email&utm_medium=Signature&utm_campaign=Free%20available
>>> 
>> 

Reply via email to