So, what's the next step?

On Thu, Mar 15, 2018 at 12:27 AM, Dmitry Pavlov <dpavlov....@gmail.com> wrote:
> Yes, I think I could move IgniteReproducingSuite to dev-utils module later.
> Thank you for this idea.
>
> Yes, It is probably it was Queries test flaky'ness.
>
> I hope Vladimir, you will find some time to make query tests more stable. It
> is not friendly to community members if their patches are rejected by
> reasons not related to their change.
>
> Any assistance from the rest of community here is also appreciated.
>
> ср, 14 мар. 2018 г. в 22:24, Vyacheslav Daradur <daradu...@gmail.com>:
>>
>> Thank you for the advice!
>>
>> Unfortunately, *IgniteReproducingSuite* is in the core module while
>> *IgniteSqlSplitterSelfTest* in the ignite-indexing module that means I
>> am not able to add the test in this test suite without addition
>> cycling dependency.
>>
>> I'd recommend you detaching *IgniteReproducingSuite* as a separate
>> module in the project to include the test suites from any module in
>> the project.
>>
>>
>> But I've prepared *Ignite Queries* in the same way as you suggested in
>> *IgniteReproducingSuite* [1] and ran all tests in
>> *IgniteSqlSplitterSelfTest* 100 times [2].
>>
>> >> IgniteBinaryCacheQueryTestSuite:
>> >> IgniteSqlSplitterSelfTest.testReplicatedTablesUsingPartitionedCacheSegmentedClient
>> >> (fail rate 0,0%)
>> For this test "Green lite" 100 times of 100.
>>
>> Green lite for all tests in *IgniteSqlSplitterSelfTest* in the latest
>> build of main PR [3].
>>
>>
>> [1]
>> https://github.com/daradurvs/ignite/blob/fd6abc915838599c2ebab3f803f90f2e641e8892/modules/indexing/src/test/java/org/apache/ignite/testsuites/IgniteCacheQuerySelfTestSuite.java
>> [2] https://ci.ignite.apache.org/viewLog.html?buildId=1136780
>> [3] https://ci.ignite.apache.org/viewLog.html?buildId=1136685
>>
>> On Wed, Mar 14, 2018 at 7:55 PM, Dmitry Pavlov <dpavlov....@gmail.com>
>> wrote:
>> > It is possible that test is failing only on agents and is always
>> > successfull
>> > locally.
>> >
>> > For researching such test there was "Ignite reproducing suite"
>> > introduced
>> > early. This suite intentionally left blank on TC. Correspondent suite in
>> > code is IgniteReproducingSuite.
>> >
>> > You may add some extra debug info into test. Add this test in
>> > IgniteReproducingSuite in code and then start suite on TC several times.
>> >
>> > ср, 14 мар. 2018 г. в 19:42, Vyacheslav Daradur <daradu...@gmail.com>:
>> >>
>> >> Dmitry, as I've written here before: I checked this test locally, many
>> >> times (didn't have any falling on 100 starts).
>> >>
>> >> On Wed, Mar 14, 2018 at 7:31 PM, Dmitry Pavlov <dpavlov....@gmail.com>
>> >> wrote:
>> >> > Hi, I've found test which never failed on master, but fails in branch
>> >> >
>> >> >  Ignite Queries [ tests 1 ]
>> >> >
>> >> >  IgniteBinaryCacheQueryTestSuite:
>> >> >
>> >> >
>> >> > IgniteSqlSplitterSelfTest.testReplicatedTablesUsingPartitionedCacheSegmentedClient
>> >> > (fail rate 0,0%)
>> >> >
>> >> >
>> >> > ср, 14 мар. 2018 г. в 19:26, Dmitry Pavlov <dpavlov....@gmail.com>:
>> >> >>
>> >> >> Hi, let me check TC run
>> >> >>
>> >> >> вт, 13 мар. 2018 г. в 9:22, Vyacheslav Daradur
>> >> >> <daradu...@gmail.com>:
>> >> >>>
>> >> >>> Dmitry,
>> >> >>>
>> >> >>> Nickolay accepted PR changes at Upsource [1].
>> >> >>>
>> >> >>> Latest ci.build [2] looks good in comparison with master [3].
>> >> >>>
>> >> >>> Following tests passed locally:
>> >> >>> CacheAffinityCallSelfTest.testAffinityCallFromClientRestartNode
>> >> >>> CacheAffinityCallSelfTest.testAffinityCallRestartNode
>> >> >>>
>> >> >>> IgniteOptimisticTxSuspendResumeMultiServerTest.testTxTimeoutOnSuspend
>> >> >>>
>> >> >>>
>> >> >>>
>> >> >>> IgniteSqlSplitterSelfTest.testReplicatedTablesUsingPartitionedCacheSegmentedClient
>> >> >>>
>> >> >>>
>> >> >>> [1] https://reviews.ignite.apache.org/ignite/review/IGNT-CR-509
>> >> >>> [2] https://ci.ignite.apache.org/viewLog.html?buildId=1134466
>> >> >>> [3] https://ci.ignite.apache.org/viewLog.html?buildId=1134372
>> >> >>>
>> >> >>> On Mon, Mar 5, 2018 at 7:16 PM, Vyacheslav Daradur
>> >> >>> <daradu...@gmail.com>
>> >> >>> wrote:
>> >> >>> > Dmitry, I saw them, but it looks like just randomness.
>> >> >>> >
>> >> >>> > I've checked it locally several times.
>> >> >>> > They failed only in one TeamCity's build of four.
>> >> >>> >
>> >> >>> > Started build once again to be sure.
>> >> >>> >
>> >> >>> > On Mon, Mar 5, 2018 at 6:59 PM, Dmitry Pavlov
>> >> >>> > <dpavlov....@gmail.com>
>> >> >>> > wrote:
>> >> >>> >> I can see Nikolay Izhikov as reviewer in Upsource.
>> >> >>> >>
>> >> >>> >> Nikolay, would you run review first?
>> >> >>> >>
>> >> >>> >> I've found several suspicious tests : Test fail rate is less
>> >> >>> >> than
>> >> >>> >> 1%,
>> >> >>> >> it is
>> >> >>> >> probably new failure
>> >> >>> >> IgniteCacheTestSuite2:
>> >> >>> >>
>> >> >>> >>
>> >> >>> >>
>> >> >>> >> GridCachePartitionedTxSingleThreadedSelfTest.testOptimisticReadCommittedRollback
>> >> >>> >> (fail rate 0,0%)
>> >> >>> >> IgniteCacheTestSuite2:
>> >> >>> >>
>> >> >>> >>
>> >> >>> >>
>> >> >>> >> GridCachePartitionedTxSingleThreadedSelfTest.testOptimisticRepeatableReadRollback
>> >> >>> >> (fail rate 0,0%)
>> >> >>> >> IgniteCacheTestSuite2:
>> >> >>> >>
>> >> >>> >>
>> >> >>> >>
>> >> >>> >> GridCachePartitionedTxSingleThreadedSelfTest.testPessimisticReadCommittedCommit
>> >> >>> >> (fail rate 0,0%)
>> >> >>> >> IgniteCacheTestSuite2:
>> >> >>> >>
>> >> >>> >>
>> >> >>> >>
>> >> >>> >> GridCachePartitionedTxSingleThreadedSelfTest.testPessimisticReadCommittedRollback
>> >> >>> >> (fail rate 0,0%)
>> >> >>> >> IgniteCacheTestSuite2:
>> >> >>> >>
>> >> >>> >>
>> >> >>> >>
>> >> >>> >> GridCachePartitionedTxSingleThreadedSelfTest.testPessimisticSerializableCommit
>> >> >>> >> (fail rate 0,0%)
>> >> >>> >>
>> >> >>> >> Vyacheslav, could you please check if these failures are related
>> >> >>> >> to
>> >> >>> >> the new
>> >> >>> >> changes?
>> >> >>> >>
>> >> >>> >>
>> >> >>> >> пн, 5 мар. 2018 г. в 18:50, Vyacheslav Daradur
>> >> >>> >> <daradu...@gmail.com>:
>> >> >>> >>
>> >> >>> >>> I've done some test-builds iteration on the weekends.
>> >> >>> >>>
>> >> >>> >>> Tests [1] look well.
>> >> >>> >>>
>> >> >>> >>> Does anyone have time to do the final review [2][3] and merge
>> >> >>> >>> it?
>> >> >>> >>>
>> >> >>> >>>
>> >> >>> >>> [1] https://ci.ignite.apache.org/viewLog.html?buildId=1125676
>> >> >>> >>> [2] https://github.com/apache/ignite/pull/3578
>> >> >>> >>> [3] https://reviews.ignite.apache.org/ignite/review/IGNT-CR-509
>> >> >>> >>>
>> >> >>> >>>
>> >> >>> >>> On Fri, Mar 2, 2018 at 10:17 PM, Vyacheslav Daradur
>> >> >>> >>> <daradu...@gmail.com>
>> >> >>> >>> wrote:
>> >> >>> >>> > Hi, Igniters!
>> >> >>> >>> >
>> >> >>> >>> > This task [1] is about 'get' requests distribution between
>> >> >>> >>> > primary
>> >> >>> >>> > and
>> >> >>> >>> > backup nodes in the replicated cache if 'readFromBackup' flag
>> >> >>> >>> > is
>> >> >>> >>> > enabled.
>> >> >>> >>> >
>> >> >>> >>> > I've prepared a solution [2] suggested by Alexei Scherbakov
>> >> >>> >>> > in
>> >> >>> >>> > Jira
>> >> >>> >>> > comments. It passed prereviews by Alexei and Nikolay Izhikov.
>> >> >>> >>> >
>> >> >>> >>> > TeamCity tests look similar with the master branch.
>> >> >>> >>> >
>> >> >>> >>> > Could someone of core module maintainers do the final review
>> >> >>> >>> > [2][3]?
>> >> >>> >>> >
>> >> >>> >>> >
>> >> >>> >>> > [1] https://issues.apache.org/jira/browse/IGNITE-5357
>> >> >>> >>> > [2] https://github.com/apache/ignite/pull/3578
>> >> >>> >>> > [3]
>> >> >>> >>> > https://reviews.ignite.apache.org/ignite/review/IGNT-CR-509
>> >> >>> >>> >
>> >> >>> >>> > --
>> >> >>> >>> > Best Regards, Vyacheslav D.
>> >> >>> >>>
>> >> >>> >>>
>> >> >>> >>>
>> >> >>> >>> --
>> >> >>> >>> Best Regards, Vyacheslav D.
>> >> >>> >>>
>> >> >>> >
>> >> >>> >
>> >> >>> >
>> >> >>> > --
>> >> >>> > Best Regards, Vyacheslav D.
>> >> >>>
>> >> >>>
>> >> >>>
>> >> >>> --
>> >> >>> Best Regards, Vyacheslav D.
>> >>
>> >>
>> >>
>> >> --
>> >> Best Regards, Vyacheslav D.
>>
>>
>>
>> --
>> Best Regards, Vyacheslav D.



-- 
Best Regards, Vyacheslav D.

Reply via email to