;
> >>>> Sure, I don't expect it to be free.
> >>>> But everybody should be aware of the cost of adding this code style,
> i.e.,
> >>>> spending a huge amount of time on reformatting and documenting code.
> >>>>
> &g
-1 from my side. I just found a serious issue with the KryoSerializer
(FLINK-2800) which in some cases produced duplicated elements or corrupted
data. I opened a PR to fix the issue (
https://github.com/apache/flink/pull/1308).
Cheers,
Till
On Wed, Oct 28, 2015 at 5:38 PM, Maximilian Michels w
What does the log of the failed TaskManager 10.0.88.140 say?
On Thu, Oct 29, 2015 at 6:44 PM, Greg Hogan wrote:
> I removed the use of numactl but left in starting two TaskManagers and am
> still seeing TaskManagers crash.
> From the JobManager log:
>
> 17:36:06,412 WARN
> akka.remote.ReliableDe
gt; (used/committed/max)]
> > > > 17:35:51,795 INFO
> > > > org.apache.flink.runtime.taskmanager.TaskManager -
> Garbage
> > > > collector stats: [PS Scavenge, GC TIME (ms): 797, GC COUNT: 5], [PS
> > > > MarkSweep, GC TIME (ms): 974, G
Hi Santosh,
how do you generate your job jar. The second error with the
`NoClassDefFoundError` usually happens when not all required runtime
classes are shipped to the cluster (either not included in the fat jar or
not explicitly added in case of a non-fat jar).
Cheers,
Till
On Mon, Nov 2, 2015
I fear that I have to give another -1.
The reason is that in the current RC Flink fails when spilling partitions
of the MutableHashTable when performing a join operation with many records
which are also large in size. Thus, effectively the join operation is
broken.
The underlying problem is that
Could you please open or update the corresponding JIRA issue if existing.
On Wed, Nov 4, 2015 at 11:14 AM, Gyula Fóra wrote:
> Hey,
>
> I found an interesting failure in the KafkaITCase, I am not sure if this
> happened before.
>
> It received a duplicate record and failed on that (not the usual
Great, thanks :-)
On Wed, Nov 4, 2015 at 12:07 PM, Maximilian Michels wrote:
> Thanks Gyula. Here is the issue:
> https://issues.apache.org/jira/browse/FLINK-2965
>
> On Wed, Nov 4, 2015 at 12:00 PM, Gyula Fóra wrote:
> > done
> >
> > Till Rohrmann ezt írta
The error should also be in 0.10. The problem is that due to the latest
stream refactorings we not only have fold operations which are
StreamOperators but also WindowFunctions (if you apply them on windows).
For the StreamOperators the output type is correctly set in the StreamGraph
and thus the in
, 15:23 Maximilian Michels wrote:
>
> > That's basically what I also found out too so far. If you want to fix
> > it please go ahead.
> >
> > On Wed, Nov 4, 2015 at 3:11 PM, Till Rohrmann
> > wrote:
> > > The error should also be in 0.10. The proble
I agree that it would be nice to be able to simply copy paste examples from
Flink's website. But it would also be nice if one could hide the imports
because they usually take a lot of space.
Cheers,
Till
On Wed, Nov 18, 2015 at 12:09 PM, Robert Metzger
wrote:
> Hi,
>
> I helped somebody yesterd
@Vasia, do you mean FLINK-3013 or FLINK-3012?
Will merge PRs for FLINK-3036 and FLINK-3013 this afternoon.
On Wed, Nov 18, 2015 at 1:47 PM, Stephan Ewen wrote:
> +1 for a timely 0.10.1 release
>
> I would like to add FLINK-2974 - periodic kafka offset committer for case
> where checkpointing is
; On Wed, Nov 18, 2015 at 2:35 PM, Ufuk Celebi wrote:
>
> > @Suneel: I think that's OK for the next major release :)
> >
> > On Wed, Nov 18, 2015 at 2:17 PM, Suneel Marthi
> > wrote:
> >
> > > Would be nice to have Flink-2949 in the mix, but I won'
There are other related PRs pending.
>
> On Thu, Nov 19, 2015 at 3:42 AM, Till Rohrmann
> wrote:
>
> > Yes forgot about merging them to release-0.10. Will do it right away.
> > On Nov 19, 2015 9:40 AM, "Robert Metzger" wrote:
> >
> > > @Suneel, I th
019, 3028 Cancel jobs in RESTARTING state
> > >
> > >
> > > Is anybody volunteering to be a release manager?
> > >
> > >
> > > On Thu, Nov 19, 2015 at 10:16 AM, Suneel Marthi
> > > wrote:
> > >
> > > > Flink-3017, Fli
es with the current release?
> > >
> > > To me it feels a little bit like rushing this out and we will have
> almost
> > > the same situation afterwards.
> > >
> > > I might be wrong but I think people should get a chance to try this
> out.
> > >
ents
> only? I would have a fix for a bug in Storm compatibility -- not sure if
> it should be included or not
>
> -Matthias
>
> On 11/20/2015 12:35 PM, Till Rohrmann wrote:
> > The optimizer bug (https://issues.apache.org/jira/browse/FLINK-3052)
> should
> > be
Hi Andre,
what you could do is to either add a new startCluster method which takes as
a parameter the task manager memory size. Alternatively, and I would prefer
this alternative, you could change the startCluster(Configuration,
StreamingMode, boolean) method to always overwrite the configuration
If not API breaking before 1.0, then probably never?
On Tue, Nov 24, 2015 at 3:06 PM, Stephan Ewen wrote:
> I was also thinking of deprecating that. With that, RichFunctions should
> change "open(Configuration)" --> "open()".
>
> Would be heavily API breaking, so bit hesitant there...
>
> On Tue
Checked checksums for src release and Hadoop 2.7 Scala 2.10 release
Checked binaries in source release
- contains ./flink-staging/flink-avro/src/test/resources/testdata.avro
License
- no new files added which are relevant for licensing
Build Flink and run tests from source release for Hadoop 2.5
Alright, then I withdraw my remark concerning testdata.avro.
On Wed, Nov 25, 2015 at 2:56 PM, Stephan Ewen wrote:
> @Till I think the avro test data file is okay, the "no binaries" policy
> refers to binary executables, as far as I know.
>
> On Wed, Nov 25, 2015 at
Hi Martin,
I think the problem is that the the WritableSerializer, WritableComparator,
ValueComparator, ValueSerializer and the AvroSerializer all use Kryo to
copy objects. However, in some cases, e.g. missing no-arg constructor, Kryo
is not able to copy the object. In these cases, one should try
The issue is https://issues.apache.org/jira/browse/FLINK-3088.
On Fri, Nov 27, 2015 at 12:21 PM, Till Rohrmann
wrote:
> Hi Martin,
>
> I think the problem is that the the WritableSerializer, WritableComparator,
> ValueComparator, ValueSerializer and the AvroSerializer all use Kr
the members of that class could be
> serialized by Flink's own serialization mechanism?
>
> Again, thanks.
>
> Best,
> Martin
>
> On 27.11.2015 12:36, Till Rohrmann wrote:
> > The issue is https://issues.apache.org/jira/browse/FLINK-3088.
> >
> > On F
Thanks Robert for being the release manager for 0.10.1
On Fri, Nov 27, 2015 at 4:21 PM, Maximilian Michels wrote:
> Great. We released that one fast. Thanks Robert.
>
> On Fri, Nov 27, 2015 at 3:27 PM, Robert Metzger
> wrote:
> > The Flink PMC is pleased to announce the availability of Flink 0.
I checked the code and the MapFunction InitializeWorkSet contains indeed
two instances of type Either. Since we use Java serialization to ship the
operators to the cluster, the MapFunction has to be serializable. Either
you make the Either type serializable or you create these instances in the
open
If I'm not mistaken, then the scheduler has already a preference to spread
independent pipelines out across the cluster. At least he uses a queue of
instances from which it pops the first element if it allocates a new slot.
This instance is then appended to the queue again, if it has some resources
Hi Sachin,
I can understand your dissatisfaction with the review process of your ML
PRs. They are open for a long time without much activity from a committer
even though you've spent a lot of effort writing them. I'm sorry for this,
since it's probably mainly because of me lacking the time resourc
+1 from my side as well. Good idea.
On Thu, Dec 10, 2015 at 11:00 PM, jun aoki wrote:
> The pluggable architecture is great! (why Don't we call it Flink plugins?
> my 2 cents)
> It will be nice if we come up with an idea of what directory structure
> should look like before start dumping connect
The only thing I did was to use the Py4J bridge to ship data between the
Scala and the Python process. That way, I could visualize data coming from
Flink with Python. I couldn't run Flink programs using the Python API,
though.
@Gyula, you can find the code here:
https://github.com/tillrohrmann/inc
Happy New Year :-) Hope everyone had a great start into the new year.
On Thu, Dec 31, 2015 at 12:57 PM, Slim Baltagi wrote:
> Happy New Year to you and your families!
> Let’s make 2016 the year of Flink: General Availability, faster growth,
> wider industry adoption, …
> Slim Baltagi
> Chicago,
Hi,
yes, initially we thought about introducing a LabeledVector where the label
can be a vector. However, for the sake of simplicity we decided to first
implement a LabeledVector with a single double value as label.
A simple double value should take 8 bytes of memory space. The
DenseVector(Array(
Hi Arjun,
welcome to the Flink community :-)
On Thu, Jan 7, 2016 at 5:40 AM, Arjun Rao wrote:
> Hi,
>
> I am new to Apache Flink and I really like the look of the API. I have been
> working with Storm for the past year and have questions about the
> DataStream API among others.
>
> 1. What are
Hi everybody,
recently we've seen an increased interest in complex event processing (CEP)
by Flink users. Even though most functionality is already there to solve
many use cases it would still be helpful for most users to have an easy to
use library. Having such a library which allows to define co
+1 since it increase maintainability of the code base if it is not really
used and thus removed.
On Fri, Jan 8, 2016 at 5:33 PM, Ufuk Celebi wrote:
> +1
>
> I wanted to make a similar proposal.
>
> – Ufuk
>
> > On 08 Jan 2016, at 17:03, Kostas Tzoumas wrote:
> >
> > for clarification, I was tal
to keep track of this proposed new feature?
>
> We should create one to keep track updates on the effort.
>
> Thanks,
>
> Henry
>
> On Fri, Jan 8, 2016 at 6:54 AM, Till Rohrmann <mailto:trohrm...@apache.org>> wrote:
> Hi everybody,
>
> recently we've see
First of all, it's a great design document. Looking forward having stream
SQL in the foreseeable future :-)
I think it is a good idea to consolidate stream SQL and CEP in the long
run. CEP's additional features compared to SQL boil down to pattern
detection. Once we have this, it should be only a
+1 for protecting the master branch.
On Wed, Jan 13, 2016 at 11:42 AM, Li, Chengxiang
wrote:
> +1 on the original style.
> Master branch disable force pushing in case of misusing and feature branch
> enable force pushing for flexible developing.
>
> -Original Message-
> From: Gyula Fóra
Hi Fabian,
thanks for bringing this issue up. I agree with you that it would be nice
to remove the Combinable annotation if it is not really needed. Now if
people are not aware of the Combinable interface then they might think that
they are actually using combiners by simply implementing CombineFu
Hi Nick,
I think the reason is to set some common settings for all flink shaded
hadoop child modules in flink-shaded-hadoop if I’m not mistaken. But this
might also be a historic artifact.
Cheers,
Till
On Fri, Jan 15, 2016 at 12:48 AM, Nick Dimiduk wrote:
> Hi folks,
>
> I noticed today that
This might be related to the problem mentioned in
https://issues.apache.org/jira/browse/FLINK-3241 and
http://apache-flink-user-mailing-list-archive.2336050.n4.nabble.com/flink-1-0-SNAPSHOT-scala-2-11-compilation-error-td4303.html
.
Cheers,
Till
On Sun, Jan 17, 2016 at 6:26 PM, Prez Cannady
wrot
A problem is that all the BlobCache/Server tests don't extend the
`TestLogger` class which logs the names of the different tests. Maybe we
should that add as well.
On Fri, Jan 15, 2016 at 3:32 PM, Prez Cannady
wrote:
> I’ve included the full log of a test run against flink-runtime here:
>
> http
Hi Hilmi,
I think in your case it makes sense to define a custom vector of strings.
The easiest implementation could be an Array[String] or List[String].
The reason why it does not make so much sense to make Vector and DenseVector
generic is that these types are algebraic data types. How would yo
Didn't we just had the discussion in another email thread?
On Jan 18, 2016 8:55 PM, "Hilmi Yildirim" wrote:
> Hi,
> the Vector and DenseVector implementations of Flink ML only allow Double
> values. But there are cases where the values are not Doubles, e.g. in NLP.
> Does it make sense to make th
> > >>>>
> > >>>> +1
> > >>>>
> > >>>> On Wed, Jan 13, 2016 at 12:37 PM, Matthias J. Sax >
> > wrote:
> > >>>>
> > >>>>> +1
> > >>>>>
> > >>>&
Hi Hilmi,
I've never seen such a problem. Could you please provide a bit more
background information. Which types are you serializing? Which serializer
are you using? What do you mean with recursion? Maybe you can provide a
simple example program which reproduces the problem.
The thing is that if
e(FieldSerializer.java:495)
> at com.esotericsoftware.kryo.Kryo.writeObject(Kryo.java:523)
> at
> com.esotericsoftware.kryo.serializers.ObjectField.write(ObjectField.java:61)
> at
> com.esotericsoftware.kryo.serializers.FieldSerializer.write(FieldSerializer.java:495)
> at com.esotericsoft
Hi Trevor,
at the moment we don't have a deep learning library implemented on top of
Flink. As far as I know, Suneel (smar...@apache.org) tinkered with the idea
to implement deep learning with Flink. Maybe he can tell you more about his
progress so far.
Supporting deep learning would be awesome :
+1
On Mon, Jan 25, 2016 at 3:36 PM, Stephan Ewen wrote:
> +1
>
> On Mon, Jan 25, 2016 at 11:41 AM, Ufuk Celebi wrote:
>
> >
> > > On 25 Jan 2016, at 11:39, Maximilian Michels wrote:
> > >
> > > I won't have the time to finish the refactoring. Also, it will be
> > > pretty painful with all the
Hi Hilmi,
could you check what happened on the JobManager side?
Cheers,
Till
On Mon, Feb 1, 2016 at 2:39 PM, Hilmi Yildirim
wrote:
> Hi,
>
> I get the following exception when I execute a code similar to the
> ALSITSuite. I train a ALS model and when the following code is executed I
> get t
ere logs of the jobmanager in local mode?
>
> Best Regards,
> Hilmi
>
>
> Am 01.02.2016 um 14:50 schrieb Till Rohrmann:
>
>> Hi Hilmi,
>>
>> could you check what happened on the JobManager side?
>>
>> Cheers,
>> Till
>>
>>
>
+1, since user were asking for it.
On Tue, Feb 2, 2016 at 10:50 AM, Fabian Hueske wrote:
> +1
>
> 2016-02-02 10:44 GMT+01:00 Maximilian Michels :
>
> > Hi Ufuk,
> >
> > +1 If the fixes are straightforward such that we don't need to test
> > extensively, I'm all for it! Releasing doesn't take muc
I like the idea to support partial functions with Flink’s Scala API.
However, I think that breaking the API and making it inconsistent with
respect to the Java API is not the best option. I would rather be in favour
of the first proposal where we add a new method xxxWith via implicit
conversions.
ensions.AcceptPartialFunctions).
>
> On Mon, Feb 8, 2016 at 2:51 PM, Till Rohrmann
> wrote:
>
> > I like the idea to support partial functions with Flink’s Scala API.
> > However, I think that breaking the API and making it inconsistent with
> > respect to the Java API is n
d join don't look like they take a function.
>
> On Tue, Feb 9, 2016 at 10:08 AM, Till Rohrmann
> wrote:
>
> > This looks like a good design to me :-) The only thing is that it is not
> > complete. For example, the filter, mapPartition, coGroup and join
> functio
lbit.io> wrote:
> >
> > > I see, thanks for the tip! I'll work on it; meanwhile, I've added some
> > > functions and Scaladoc:
> > >
> > >
> >
> https://github.com/radicalbit/flink/blob/1159-implicit/flink-scala/src/main/scala/org/apache/flink/a
Thanks for being our release manager Ufuk :-) Great work!
On Fri, Feb 12, 2016 at 2:15 PM, Robert Metzger wrote:
> Thank you for doing a release Ufuk!
>
> I just tweeted about it:
> https://twitter.com/ApacheFlink/status/698130110709428224
>
>
> On Fri, Feb 12, 2016 at 2:13 PM, Maximilian Michel
I agree with Martin that the original topic of this thread was about how to
keep FlinkML active so that new changes will be promptly merged. The things
we want to implement is then up to the active contributors of FlinkML.
Personally, I would prefer to keep FlinkML as part of Flink's main
reposito
I think the important part about the ConfigConstants is that the values
don’t change. How they are represented inside of Flink, does not really
matter. It would be good if that could be verified automatically.
Cheers,
Till
On Tue, Feb 16, 2016 at 2:59 PM, Robert Metzger wrote:
> Thank you for
Hi Simone,
as far as I know, there is nobody currently working on an H2O integration.
I only looked briefly at the sparkling water implementation when it was
released. If I remember correctly, then the general idea was to start H2O
from within the Executor thread and to use a special RDD to commun
Judging from our chaining condition
ds.getPushChainDriverClass() != null &&
!(pred instanceof NAryUnionPlanNode) &&// first op after union is
stand-alone, because union is merged
!(pred instanceof BulkPartialSolutionPlanNode) &&// partial
solution merges anyways
!(pred instanceof WorksetPl
The idea was actually to leverage existing linear algebra libraries such as
breeze instead of building another blas implementation which will never be
as good as the ones out there.
Cheers,
Till
On Fri, Feb 19, 2016 at 9:48 AM, Theodore Vasiloudis <
theodoros.vasilou...@gmail.com> wrote:
> Just
Furthermore, if all of your sources are stoppable, then you can even stop
the streaming job gracefully now either using the CLI or the StopJob
message.
Cheers,
Till
On Fri, Feb 19, 2016 at 12:16 PM, Márton Balassi
wrote:
> Adding to Ufuk's answer: yes, cancelling the job frees up the resource
I just tested executing a streaming example on the current master and
everything worked. Can you try clearing the IntelliJ cache and rebuild the
project?
Cheers,
Till
On Thu, Feb 25, 2016 at 5:13 PM, Tara Athan wrote:
> Hi, I am just exploring Flink, and have run into a curious issue. I have
>
Hi Timothy,
you can use the `JDBCOutputFormat` in combination with the
`DataStream.writeUsingOutputFormat` method. However, this won't give you
exactly once guarantees since the output formats don't take part in the
checkpointing mechanism.
Currently, Chesnay is working on a generic framework for
Hi Vasia,
that is because you're missing the flink-gelly dependency. If you just
build flink-gelly-examples, then it won't contain flink-gelly because it is
not a fat jar. You either have to install flink-gelly on your cluster or
package it in the final user jar.
Cheers,
Till
On Sat, Feb 27, 201
; to examples and cluster execution instructions are broken.
> I'll create an issue.
>
> -V.
>
> On 29 February 2016 at 11:54, Till Rohrmann wrote:
>
> > Hi Vasia,
> >
> > that is because you're missing the flink-gelly dependency. If you just
> >
s. I am able to run the streaming examples after
> invalidating caches, restarting and rebuild project.
>
> BTW I see a message that says 100 errors during compilation. Is that to be
> expected?
>
> Tara
>
> On 2/26/16 4:02 AM, Till Rohrmann wrote:
>
>> I just tested
If possible I would like to include
https://issues.apache.org/jira/browse/FLINK-3557
On Tue, Mar 1, 2016 at 3:29 PM, Robert Metzger wrote:
> Okay. I'll create a new RC today.
>
> Anything else we want to get into the RC?
>
> On Tue, Mar 1, 2016 at 3:26 PM, Ufuk Celebi wrote:
>
> > -1
> >
> > Un
Great to hear that you two are giving a talk at ApacheCon.
As far as I know, there's nobody working on a streaming interpreter for
Zeppelin. People thought about doing it, but it never got realized so far.
But I think it should not be too difficult to implement. So if you wanna
take the lead there
pelin, which is
> possible by finding out the job-id and sending a cancel message to the task
> manager.
> I'll continue integrating it with the current versions of Flink and
> Zeppelin as soon as I have the time.
>
> cheers,
> Nik
>
>
> 2016-03-02 10:27 GMT+01:00
1 14:31:26 2016 +0100
>
> [FLINK-3554] [streaming] Emit a MAX Watermark after finite sources
> finished
>
> This closes #1750
>
> commit 8949ccf66b211b3c5cd8e66557afbff21fb093a6
> Author: Till Rohrmann
> Date: Tue Mar 1 12:36:22 2016 +0100
>
> [FLINK-355
+1
Checked that the sources don't contain binaries
Tested cluster execution with flink/run and web client job submission
Run all examples via FliRTT
Tested Kafka 0.9
Verified that quickstarts work with Eclipse and IntelliJ
Run example with RemoteEnvironment
Verified SBT quickstarts
On Thu, Mar 3,
I agree with Aljoscha on this one, because `DataStreamSink` only contains
setters which are compatible with the Scala API.
On Mon, Mar 14, 2016 at 11:02 AM, Aljoscha Krettek
wrote:
> By the way, I don’t think it’s a bug that addSink() returns the Java
> DataStreamSink. Having a Scala specific ve
Hi Tianqi,
dmlc looks really cool and it would be great to integrate it with Flink. As
far as I understood your requirements, I think that you can already
implement most of it on Flink.
For example, starting a special container which does not receive any input
could be a specialized SourceOperato
Hi Greg,
could you share an example program with us which reproduces the problem? I
suspect that, somehow, your user code class BlockInfo is sent directly to
the JobManager where it is deserialized without the user code class loader.
Cheers,
Till
On Tue, Mar 15, 2016 at 4:19 PM, Greg Hogan wr
Mar 15, 2016 at 5:16 PM, Greg Hogan wrote:
>
> > Hi Till,
> >
> > The code in question is part of FLINK-2909 which is currently running on
> > Travis but which does not trigger this issue. I'll keep looking.
> >
> > Thanks,
> > Greg
> >
> >
Hi Dmitriy,
I’m not sure whether I’ve understood your question correctly, so please
correct me if I’m wrong.
So you’re asking whether it is a problem that
stat1 = A.map.reduce
A = A.update.map(stat1)
are executed on the same input data set A and whether we have to cache A
for that, right? I ass
+1
On Wed, Mar 23, 2016 at 11:24 AM, Stephan Ewen wrote:
> Yes, there is also the Rich Scala Window Functions, and the tests that used
> to address wrong JAR directories.
>
> On Wed, Mar 23, 2016 at 11:15 AM, Ufuk Celebi wrote:
>
> > Big +1, let's get this rolling... ;)
> >
> > On Wed, Mar 23,
Hi Trevor,
great to hear that you have a working prototype :-) And it is also good
that you shared your insights you gained when implementing it. Flink’s ML
library is far from perfect and, thus, all kinds of feedback is highly
valuable. In general it is always good to contribute code back if you
e construct such a loop with a
> concrete
> > > example?
> > >
> > > Let's take something nonsensical yet specific.
> > >
> > > Say, in samsara terms we do something like that :
> > >
> > > var avg = Double.PositiveInfinit
Hi Sergio,
could you please provide a complete example (including input data) to
reproduce your problem. It is hard to tell what's going wrong when one only
sees a fraction of the program.
Cheers,
Till
On Tue, Mar 29, 2016 at 5:58 PM, Sergio Ramírez
wrote:
> Hi again,
>
> I've not been able to
possible, and the algorithms that need it, just will
> > have to work around this. I think this answers my question -- thanks!
> >
> > -d
> >
> >
> > On Tue, Mar 29, 2016 at 2:53 AM, Till Rohrmann
> > wrote:
> >
> >> Hi,
> >>
>
I would assume that something went wrong on the JobManager side. Could you
check the logs if they contain something suspicious? Additionally you could
turn on lifecycle event logging for Akka.
Cheers,
Till
Hi Do,
the easiest way is to avoid using methods which trigger an eager execution
(collect, count, print) but to define sinks instead. Alternatively, you can
persist intermediate results by writing them to disk and continue
processing from there. That way, you won't re-calculate all parts of your
Could you please rerun the whole job with debug log level and logging of
Akka's lifecycle events turned on?
Cheers,
Till
On Thu, Mar 31, 2016 at 4:21 PM, Matthias J. Sax wrote:
> enclosed the logs.. maybe you can make some sense out if them.
>
> On 03/31/2016 02:52 PM, Till
deserialize(Serialization.scala:98)
> > >>> at
> > >> akka.remote.MessageSerializer$.deserialize(MessageSerializer.scala:23)
> > >>> at
> > >>
> >
> akka.remote.DefaultMessageDispatcher.payload$lzycompute$1(Endpoint.scala:58)
&
Hi Flink community,
I've written a short blog [1] post about Flink's new CEP library which
basically showcases its functionality using a monitoring example. I would
like to publish the post on the flink.apache.org blog next week, if nobody
objects. Feedback is highly appreciated :-)
[1]
https://d
Hi Flink community,
I recently started working on dynamic scaling. As a first step we want to
introduce state sharding which is a requirement for partitioned state to be
re-distributable. The following design doc [1] explains the ideas behind
it. Any feedback is highly appreciated :-)
[1]
https:/
e put it on Flink Blog
>
> Cheers
> Gen
>
>
> On Fri, Apr 1, 2016 at 9:56 PM, Till Rohrmann
> wrote:
>
>> Hi Flink community,
>>
>> I've written a short blog [1] post about Flink's new CEP library which
>> basically showcases its functionali
Sorry, my bad. Comments should work now.
On Mon, Apr 4, 2016 at 3:51 PM, Aljoscha Krettek
wrote:
> Comments are not enabled.
>
> On Mon, 4 Apr 2016 at 13:58 Till Rohrmann wrote:
>
> > Hi Flink community,
> >
> > I recently started working on dynamic scali
only work
> with 1.0.1.
>
> On Mon, Apr 4, 2016 at 3:35 PM, Till Rohrmann
> wrote:
> > Thanks a lot to all for the valuable feedback. I've incorporated your
> > suggestions and will publish the article, once Flink 1.0.1 has been
> released
> > (we need 1
Hi Greg,
I like the idea to have a macro-benchmarking suite to exactly test the
points you've mentioned. If we don't have reliable performance numbers,
then it will always be hard to tell whether an improvement makes sense or
not (performance-wise).
I think we already undertook a first attempt to
Given that it's low overhead, then +1 for new bug fix release. If we find
more stuff later on, then we can release another bug fix version.
Cheers,
Till
On Mon, Apr 11, 2016 at 6:32 PM, Ufuk Celebi wrote:
> Hey all,
>
> we recently released Flink 1.0.1, but some fixes did not make it into
> the
Hey guys,
I have a suspicion which could be the culprit: Could change the line
KryoSerializer.java:328 to kryo.setReferences(true) and try if the error
still remains? We deactivated the reference tracking and now Kryo shouldn’t
be able to resolve cyclic references properly.
Cheers,
Till
On Mon
+1
On Tue, Apr 12, 2016 at 1:13 PM, Robert Metzger wrote:
> Good catch Till!
>
> I just checked it with the Mahout source code and the issues is gone with
> reference tracking enabled.
>
> I would just re-enable it again in Flink.
>
> On Tue, Apr 12, 2016 at 10:20 AM
Hi jstar,
what's exactly the problem you're observing?
Cheers,
Till
On Wed, Apr 13, 2016 at 2:23 PM, star jlong
wrote:
> Hi there,
>
> I'm jstar. I have been playing around with flink. I'm very much interested
> in submitting a topoloy to flink using its api. As indicated
> on stackoverflow,
tp://stackoverflow.com/questions/36584784/issues-while-submitting-a-topology-to-apache-flink-using-the-flink-api
>
> On 13.04.2016 14:28, Till Rohrmann wrote:
> > Hi jstar,
> >
> > what's exactly the problem you're observing?
> >
> > Cheers,
> > T
I've updated the master. Could you check it out and run your program with
the latest master? I would expect to see a ClassNotFoundException.
On Wed, Apr 13, 2016 at 2:54 PM, Till Rohrmann wrote:
> Yes that is true. I'll commit a hotfix for that.
>
> My suspicion is that we
> indicated on that post because I'm the one that posted that issue.
> > >
> > > Le Mercredi 13 avril 2016 13h35, Chesnay Schepler <
> > ches...@apache.org> a écrit :
> > >
> > >
> > >
> >
> http://stackoverflow.com/que
501 - 600 of 1469 matches
Mail list logo