Till,
An issue with your suggestion is that the job state may grow unbounded. You
are managing
expiration of data from the cache in the operator, but the state is
partitioned by the stream key.
That means if we no longer observe a key, the state associated with that
key will never be
removed.
In
Right now I'm using Flink 1.0.2...to which version should I downgrade?
The hardware seems to be ok..how could I detect a faulty hardware?
These errors appeared in every run of my job after I moved the temporary
directory from ssd to hdd and I extended my pipeline with a dataset that
grows as the
then this.program of PackageProgram's object has null. [1]
in run() of CliFrontend.java, invokes executeProgramBlocking(program, client,
userParallelism) [2]
and this method invokes client.runBlocking(program, parallelism) [3]
and this method invokes runBlocking(prog.getPlanWithJars(),
Hello!
Using flink 1.0.2, I noticed that exceptions thrown during a flink program
would show up on the flink dashboard in the 'Exceptions' tab. That's great!
However, I don't think flink currently logs this same exception. I was
hoping there would be an equivalent `log.error` call so that third
The problem seems to occur quite often.
Did you update your Flink version recently? If so, could you try to
downgrade and see if the problem disappears.
Is it otherwise possible that it is cause by faulty hardware?
2016-05-20 18:05 GMT+02:00 Flavio Pompermaier :
> This
This time (Europed instead of Europe):
java.lang.Exception: The data preparation for task 'CHAIN GroupReduce
(GroupReduce at createResult(PassaggioWithComprAndVend.java:132)) ->
Map (Key Extractor)' , caused an error: Error obtaining the sorted
input: Thread 'SortMerger spilling thread'
Nope - thanks for advice ;) I will try that tomorrow
2016-05-20 11:24 GMT+02:00 Ufuk Celebi :
> I would suggest to set the log level to DEBUG and check the logs why
> the client can not connect to your elastic search cluster. Did you try
> that?
>
> On Mon, May 9, 2016 at 3:54
Hi folks
I'm extending a RichFlatMapFunction in order to use states on a keyed stream.
Concerning this i have two questions:
1. I have a var state_item: ValueState[Option[String]] as a local variable in
this class. Initialized with state_item = getRuntimeContext.getState(new
This time another error (rerialization instead of serialization):
com.esotericsoftware.kryo.KryoException: Unable to find class:
it.okkam.flink.entitons.*rerialization*.pojo.EntitonQuadPojo
at
com.esotericsoftware.kryo.util.DefaultClassResolver.readName(DefaultClassResolver.java:138)
Ok..great to hear that!
Thanks to all for the support
On Fri, May 20, 2016 at 3:53 PM, Ufuk Celebi wrote:
> On Fri, May 20, 2016 at 3:27 PM, Aljoscha Krettek
> wrote:
> > I think it might just be a warning. When using Kryo it is in the end a
> >
On Fri, May 20, 2016 at 3:27 PM, Aljoscha Krettek wrote:
> I think it might just be a warning. When using Kryo it is in the end a
> GenericTypeInformation but the TypeAnalyzer might still initially try to
> analyze it as a POJO.
Yes, that's correct.
To get back at the
Indeed you can test this problem with:
import org.apache.flink.api.java.ExecutionEnvironment;
import org.joda.time.DateTime;
import de.javakaffee.kryoserializers.jodatime.JodaDateTimeSerializer;
public class DateTimeError {
public static void main(String[] args) throws Exception {
Thats correct. With the fully async option the checkpoints take longer but
you don't impact ongoing processing of elements. With the semi-async method
snapshots take a shorter time but during the synchronous part no element
processing can happen.
On Fri, 20 May 2016 at 15:04 Abhishek Singh
I think it might just be a warning. When using Kryo it is in the end a
GenericTypeInformation but the TypeAnalyzer might still initially try to
analyze it as a POJO.
On Fri, 20 May 2016 at 12:51 Ufuk Celebi wrote:
> I tried to reproduce this and if you set up the depdency
>
>
Aljoscha answered this in the other thread you started for this
("'Last One' Window")
On Fri, May 20, 2016 at 12:43 PM, Artem Bogachev
wrote:
> Hi,
>
> I’ve faced a problem trying to model our platform using Flink Streams.
>
> Let me describe our model:
>
> // Stream
On Fri, May 20, 2016 at 3:12 PM, Abhishek Singh
wrote:
> Thanks. I am still in theory/evaluation mode. Will try to code this up to
> see if checkpoint will become an issue. I do have a high rate of ingest and
> lots of in flight data. Hopefully flink back pressure
Thanks. I am still in theory/evaluation mode. Will try to code this up to
see if checkpoint will become an issue. I do have a high rate of ingest and
lots of in flight data. Hopefully flink back pressure keeps this
nicely bounded.
I doubt it will be a problem for me - because even spark is
Yes. Thanks for explaining.
On Friday, May 20, 2016, Ufuk Celebi wrote:
> On Thu, May 19, 2016 at 8:54 PM, Abhishek R. Singh
> > wrote:
> > If you can take atomic in-memory copies, then it works (at the cost of
> > doubling your
You have to send this to
user-unsubscr...@flink.apache.org
You've sent this this to user@f.a.o instead.
On Fri, May 20, 2016 at 1:01 PM, Christophe Salperwyck
wrote:
>
On Thu, May 19, 2016 at 7:48 PM, Abhishek R. Singh
wrote:
> There seems to be some relationship between watermarks, triggers and
> checkpoint that is someone not being leveraged.
Checkpointing is independent of this, yes. Did the state size become a
problem for
I tried to reproduce this and if you set up the depdency
de.javakaffee
kryo-serializers
0.28
and register the Kryo type as suggested you should not see any log
messages (e.g. the type should be treated as a generic type and not a
pojo type).
Does the program
Hi Ufuk,
my records could be quite large Pojos (I think some MB).
The only thing I do to configure Kryo is:
env.registerTypeWithKryoSerializer(DateTime.class,
JodaDateTimeSerializer.class );
Best,
Flavio
On Fri, May 20, 2016 at 12:38 PM, Ufuk Celebi wrote:
> @Stefano: the
Hi,
I’ve faced a problem trying to model our platform using Flink Streams.
Let me describe our model:
// Stream of data, ex. stocks: (AAPL, 100.0), (GZMP, 100.0) etc.
val realData: DataStream[(K, V)] = env.addSource(…)
// Stream of forecasts (same format) based on some window aggregates
val
@Stefano: the records are serialized anyway for batch jobs. The
spilling deserializer is only relevant if single records are very
large. How large are your records? In any case, I don't expect this to
be the problem.
@Flavio: The class name "typo" errors (Vdhicle instead of Vehicle and
ttil
On Thu, May 19, 2016 at 4:46 PM, 윤형덕 wrote:
>
> how can this.mainClass that doesn't override getPlan method that is
abstract method of Program interface(program.class) and has only static
main method be instantiate as Program?
This is only called if the class is actually a
I think you are looking for the savepoints feature:
https://ci.apache.org/projects/flink/flink-docs-release-1.0/apis/streaming/savepoints.html
The general idea is to trigger a savepoint, start the second job from
this savepoint (reading from the same topic), and then eventually
cancel the first
Hey Arun!
How did you configure your Kafka source? If the offset has been
committed and you configured the source to read from the latest
offset, the message should not be re-processed.
– Ufuk
On Fri, May 13, 2016 at 2:19 PM, Arun Balan wrote:
> Hi, I am trying to use the
Indeed I can confirm that I resolved this problem reducing the number of
slots per Task Manager (and thus incrementing the available memory of each
task)!
However from time to time I have serialization issue that I don't know
where they come from..it looks like the PjoSerialization has some issue
The job is running out of heap memory, probably because a user
function needs a lot of it (the parquet thrift sink?).
You can try to work around it by reducing the amount of managed memory
in order to leave more heap space available.
On Thu, May 12, 2016 at 6:55 PM, Flavio Pompermaier
I would suggest to set the log level to DEBUG and check the logs why
the client can not connect to your elastic search cluster. Did you try
that?
On Mon, May 9, 2016 at 3:54 PM, rafal green wrote:
> Dear Sir or Madam,
>
> Can you tell me why I have a problem with
Today I've got this other strange error.. Obviously I don't have a
VdhicleEvent class, but a VehicleEvent class :(
java.lang.RuntimeException: Cannot instantiate class.
at
org.apache.flink.api.java.typeutils.runtime.PojoSerializer.deserialize(PojoSerializer.java:407)
at
On Thu, May 19, 2016 at 8:54 PM, Abhishek R. Singh
wrote:
> If you can take atomic in-memory copies, then it works (at the cost of
> doubling your instantaneous memory). For larger state (say rocks DB), won’t
> you have to stop the world (atomic snapshot) and make
33 matches
Mail list logo