I've ran the job once more (always using the checksum branch) and this time
I got:
Caused by: java.lang.ArrayIndexOutOfBoundsException: 1953786112
at
org.apache.flink.api.common.typeutils.base.EnumSerializer.deserialize(EnumSerializer.java:83)
at
org.apache.flink.api.common.typeutils.base.EnumSerializer.deserialize(EnumSerializer.java:32)
at
org.apache.flink.api.java.typeutils.runtime.PojoSerializer.deserialize(PojoSerializer.java:431)
at
org.apache.flink.api.java.typeutils.runtime.TupleSerializer.deserialize(TupleSerializer.java:135)
at
org.apache.flink.api.java.typeutils.runtime.TupleSerializer.deserialize(TupleSerializer.java:30)
at
org.apache.flink.runtime.io.disk.ChannelReaderInputViewIterator.next(ChannelReaderInputViewIterator.java:100)
at
org.apache.flink.runtime.operators.sort.MergeIterator$HeadStream.nextHead(MergeIterator.java:161)
at
org.apache.flink.runtime.operators.sort.MergeIterator.next(MergeIterator.java:113)
at
org.apache.flink.runtime.operators.util.metrics.CountingMutableObjectIterator.next(CountingMutableObjectIterator.java:45)
at
org.apache.flink.runtime.util.NonReusingKeyGroupedIterator.advanceToNext(NonReusingKeyGroupedIterator.java:130)
at
org.apache.flink.runtime.util.NonReusingKeyGroupedIterator.access$300(NonReusingKeyGroupedIterator.java:32)
at
org.apache.flink.runtime.util.NonReusingKeyGroupedIterator$ValuesIterator.next(NonReusingKeyGroupedIterator.java:192)
at
org.okkam.entitons.mapping.flink.IndexMappingExecutor$TupleToEntitonJsonNode.reduce(IndexMappingExecutor.java:64)
at
org.apache.flink.runtime.operators.GroupReduceDriver.run(GroupReduceDriver.java:131)
at org.apache.flink.runtime.operators.BatchTask.run(BatchTask.java:486)
at
org.apache.flink.runtime.operators.BatchTask.invoke(BatchTask.java:351)
at org.apache.flink.runtime.taskmanager.Task.run(Task.java:585)
at java.lang.Thread.run(Thread.java:745)
On Thu, Oct 6, 2016 at 11:00 AM, Ufuk Celebi <u...@apache.org> wrote:
> Yes, if that's the case you should go with option (2) and run with the
> checksums I think.
>
> On Thu, Oct 6, 2016 at 10:32 AM, Flavio Pompermaier
> <pomperma...@okkam.it> wrote:
> > The problem is that data is very large and usually cannot run on a single
> > machine :(
> >
> > On Thu, Oct 6, 2016 at 10:11 AM, Ufuk Celebi <u...@apache.org> wrote:
> >>
> >> On Wed, Oct 5, 2016 at 7:08 PM, Tarandeep Singh <tarand...@gmail.com>
> >> wrote:
> >> > @Stephan my flink cluster setup- 5 nodes, each running 1 TaskManager.
> >> > Slots
> >> > per task manager: 2-4 (I tried varying this to see if this has any
> >> > impact).
> >> > Network buffers: 5k - 20k (tried different values for it).
> >>
> >> Could you run the job first on a single task manager to see if the
> >> error occurs even if no network shuffle is involved? That should be
> >> less overhead for you than running the custom build (which might be
> >> buggy ;)).
> >>
> >> – Ufuk
> >
> >
> >
> >
>
