Hi Andrey, Did you get a chance to look into the reproducer ? thanks.
Thanks. On 11 March 2017 at 17:37, Andrey Mashenkov <andrey.mashen...@gmail.com> wrote: > Thanks Anil. > > I'll take a look. > > 11 марта 2017 г. 9:58 пользователь "Anil" <anilk...@gmail.com> написал: > > Hi Andrey, > > i am able to reproduce the issue with following reproducer. > > https://github.com/adasari/test-ignite-parallelism.git > > Please let me know if you see any issue with reproducer. > > Thanks > > On 11 March 2017 at 11:46, Anil <anilk...@gmail.com> wrote: > >> Hi Andrey, >> >> i have created test project which pushes the entries to drastreamer using >> for loop and it is working without any issue. >> >> But my application which does the same thing but loads the data from >> hbase and it is failing with *"java.lang.IllegalStateException: Data >> streamer has been closed." * >> >> >> >> *java.lang.IllegalStateException: Data streamer has been closed. >> at org.apache.ignite.internal.pro >> <http://org.apache.ignite.internal.pro>cessors.datastreamer.DataStreamerImpl.enterBusy(DataStreamerImpl.java:406) >> at org.apache.ignite.internal.pro >> <http://org.apache.ignite.internal.pro>cessors.datastreamer.DataStreamerImpl.addDataInternal(DataStreamerImpl.java:613) >> at org.apache.ignite.internal.pro >> <http://org.apache.ignite.internal.pro>cessors.datastreamer.DataStreamerImpl.addData(DataStreamerImpl.java:667)* >> >> i set the system and public thread pool as 32 for an 8 core machine and >> using 4 node cluster. >> >> and i see that loading data into cache for which queryParallelism >> configured is failing and other cache loads working fine. >> >> Thanks >> >> >> On 10 March 2017 at 05:05, Andrey Mashenkov <andrey.mashen...@gmail.com> >> wrote: >> >>> Hi Anil, >>> >>> I don't think it is a thead starvation due to it work fine for 1 thread. >>> Whould you please share a repro? >>> >>> On Thu, Mar 9, 2017 at 9:09 PM, Anil <anilk...@gmail.com> wrote: >>> >>>> >>>> Hi Andrey, >>>> >>>> i tried to set the parallelism to 2, 4 on 4 node cluster (8 core >>>> machines) and initiated the data load (using compute job). Data streamers >>>> are getting closed and data load is failing. >>>> >>>> When I set the parallelism to 1, data load working as expected. the >>>> issue could be threads starvation/ non availability of resources and not >>>> ignite issue. Do you think any other issue ? >>>> >>>> I will try on 16 core machines cluster tomorrow. >>>> >>>> Thanks. >>>> >>>> On 8 March 2017 at 18:04, Andrey Mashenkov <andrey.mashen...@gmail.com> >>>> wrote: >>>> >>>>> Hi Anil, >>>>> >>>>> SQL queries are run in system pool. >>>>> >>>>> On Wed, Mar 8, 2017 at 3:04 PM, Anil <anilk...@gmail.com> wrote: >>>>> >>>>>> Hi, >>>>>> >>>>>> Does queryparallelism uses system pool or separate thread pool other >>>>>> than system and public thread pool ? please clarify. >>>>>> >>>>>> 1. https://apacheignite.readme.io/docs/sql-performance-and-d >>>>>> ebugging#sql-performance-and-usability-considerations >>>>>> >>>>>> Thanks. >>>>>> >>>>>> >>>>> >>>>> >>>>> >>>>> -- >>>>> Best regards, >>>>> Andrey V. Mashenkov >>>>> >>>> >>>> >>> >>> >>> -- >>> Best regards, >>> Andrey V. Mashenkov >>> >> >> > >
