Hi Laurent,
Thanks for trying out Ververica platform!
However, please note that this is the mailing list of the Apache Flink
project.
Please post further questions using the "Community Edition Feedback" button
on this page: https://ververica.zendesk.com/hc/en-us
We are working on setting up a bett
Hi Marco,
You cannot really synchronize data that is being emitted via different
streams (without bringing them together in an operator).
I see two options:
1) emit the event to create the partition and the data to be written into
the partition to the same stream. Flink guarantees that records d
Hi Dongwon,
Maybe you can add your use case to the FLIP-107 discussion thread [1] and
thereby support the proposal (after checking that it would solve your
problem).
It's always helpful to learn about the requirements of users when designing
new features.
It also helps to prioritize which feature
Hi,
When running your code in the IDE, everything runs in the same local JVM.
When you run the job on Kubernetes, the situation is very different.
Your code runs in multiple JVM processes distributed in a cluster.
Flink provides a metrics collection system that you should use to collect
metrics f
Hi White,
Can you describe your problem in more detail?
* What is your Flink version?
* How do you deploy the job (application / session cluster), (Kubernetes,
Docker, YARN, ...)
* What kind of job are you running (DataStream, Table/SQL, DataSet)?
Best, Fabian
Am Mo., 20. Juli 2020 um 08:42 Uhr
Hi Brian,
AFAIK, Arvid and Piotr (both in CC) have been working on the threading
model of the checkpoint coordinator.
Maybe they can help with this question.
Best, Fabian
Am Mo., 20. Juli 2020 um 03:36 Uhr schrieb :
> Anyone can help us on this issue?
>
>
>
> Best Regards,
>
> Brian
>
>
>
> *Fr
Hi Joris,
I don't think that the approach of "add methods in operator class code that
can be called from the main Flink program" will work.
The most efficient approach would be implementing a ProcessFunction that
counts in 1-min time buckets (using event-time semantics) and updates the
metrics.
I
t;
> <https://maas.mail.163.com/dashi-web-extend/html/proSignature.html?ftlId=1&name=shadowell&uid=shadowell%40126.com&iconUrl=https%3A%2F%2Fmail-online.nosdn.127.net%2Fqiyelogo%2FdefaultAvatar.png&items=%5B%22shadowell%40126.com%22%5D>
> 签名由 网易邮箱大师 <https://mail.1
Hi Jie Feng,
As you said, Flink translates SQL queries into streaming programs with
auto-generated operator IDs.
In order to start a SQL query from a savepoint, the operator IDs in the
savepoint must match the IDs in the newly translated program.
Right now this can only be guaranteed if you transl
Congrats Yu!
Cheers, Fabian
Am Mi., 17. Juni 2020 um 10:20 Uhr schrieb Till Rohrmann <
trohrm...@apache.org>:
> Congratulations Yu!
>
> Cheers,
> Till
>
> On Wed, Jun 17, 2020 at 7:53 AM Jingsong Li
> wrote:
>
> > Congratulations Yu, well deserved!
> >
> > Best,
> > Jingsong
> >
> > On Wed, Jun
Hi Josson,
I don't have much experience setting memory bounds in Kubernetes myself,
but my colleague Andrey (in CC) reworked Flink's memory configuration for
the last release to ease the configuration in container envs.
He might be able to help.
Best, Fabian
Am Do., 21. Mai 2020 um 18:43 Uhr sch
Hi,
The code of the implementation is linked in the paper:
https://github.com/DataSystemsGroupUT/Adaptive-Watermarks
Since this is a prototype for a research paper, I'm doubtful that the
project is maintained.
I also didn't find an open-source license attached to the code.
Hence adding the project
one?
>>>>
>>>> Benchao Li 于 2020年5月5日周二 17:26写道:
>>>>
>>>>> Hi lec,
>>>>>
>>>>> You don't need to specify time attribute again like `TUMBLE_ROWTIME`,
>>>>> you just select the time attribute field
There's also the Table API approach if you want to avoid typing a "full"
SQL query:
Table t = tEnv.from("myTable");
Cheers,
Fabian
Am Di., 5. Mai 2020 um 16:34 Uhr schrieb Őrhidi Mátyás <
matyas.orh...@gmail.com>:
> Thanks guys for the prompt answers!
>
> On Tue, May 5, 2020 at 2:49 PM Kurt You
ement to make it . Can it be possible?
>
> Fabian Hueske 于2020年5月4日周一 下午4:04写道:
>
>> Hi,
>>
>> If the interval join emits the time attributes of both its inputs, you
>> can use either of them as a time attribute in a following operator because
>> the join ensur
Hi,
If the interval join emits the time attributes of both its inputs, you can
use either of them as a time attribute in a following operator because the
join ensures that the watermark will be aligned with both of them.
Best, Fabian
Am Mo., 4. Mai 2020 um 00:48 Uhr schrieb lec ssmi :
> Thanks
> }
>
> Then used a new object of GenericSerializer in the FlinkKafkaProducer
>
> FlinkKafkaProducer producer =
> new FlinkKafkaProducer<>(topic, new GenericSerializer(topic, schema,
> schemaRegistryUrl), kafkaConfig, Semantic.AT_LEAST_ONCE);
>
> Thanks , Anil.
>
>
&
Hi Sudan,
I noticed a few issues with your code:
1) Please check the computation of timestamps. Your code
public long extractAscendingTimestamp(Eventi.Event element) {
return element.getEventTime().getSeconds() * 1000;
}
only seems to look at the seconds of a timestamp. Typically, you wou
Hi Anil,
Here's a pointer to Flink's end-2-end test that's checking the integration
with schema registry [1].
It was recently updated so I hope it works the same way in Flink 1.9.
Best,
Fabian
[1]
https://github.com/apache/flink/blob/master/flink-end-to-end-tests/flink-confluent-schema-registry/
Hi,
With Flink streaming operators
However, these parts are currently being reworked to enable a better
integration of batch and streaming use cases (or hybrid use cases such as
yours).
A while back, we wrote a blog post about these plans [1]:
> *"Unified Stream Operators:* Blink extends the Fli
Hi Kristoff,
I'm not aware of any concrete plans for such a feature.
Best,
Fabian
Am So., 5. Apr. 2020 um 22:33 Uhr schrieb KristoffSC <
krzysiek.chmielew...@gmail.com>:
> Hi,
> according to [1] operator state and broadcast state (which is a "special"
> type of operator state) are not stored in
Hi Giriraj,
This looks like the deserialization of a String failed.
Can you isolate the problem to a pair of sending and receiving tasks?
Best, Fabian
Am So., 5. Apr. 2020 um 20:18 Uhr schrieb Giriraj Chauhan <
graj.chau...@gmail.com>:
> Hi,
>
> We are submitting a flink(1.9.1) job for data pro
Congrats Jingsong!
Cheers, Fabian
Am Fr., 21. Feb. 2020 um 17:49 Uhr schrieb Rong Rong :
> Congratulations Jingsong!!
>
> Cheers,
> Rong
>
> On Fri, Feb 21, 2020 at 8:45 AM Bowen Li wrote:
>
> > Congrats, Jingsong!
> >
> > On Fri, Feb 21, 2020 at 7:28 AM Till Rohrmann
> > wrote:
> >
> >> Congr
Fr., 14. Feb. 2020 um 17:48 Uhr schrieb Fabian Hueske :
> Hi everyone,
>
> We announced the program of Flink Forward San Francisco 2020.
> The conference takes place at the Hyatt Regency in San Francisco from
> March 23rd to 25th.
>
> On the first day we offer four
Hi everyone,
We announced the program of Flink Forward San Francisco 2020.
The conference takes place at the Hyatt Regency in San Francisco from March
23rd to 25th.
On the first day we offer four training sessions [1]:
* Apache Flink Developer Training
* Apache Flink Runtime & Operations Training
Congrats team and a big thank you to the release managers!
Am Mi., 12. Feb. 2020 um 16:33 Uhr schrieb Timo Walther :
> Congratualations everyone! Great stuff :-)
>
> Regards,
> Timo
>
>
> On 12.02.20 16:05, Leonard Xu wrote:
> > Great news!
> > Thanks everyone involved !
> > Thanks Gary and Yu fo
Hi,
I think you are looking for BroadcastState [1].
Best, Fabian
[1]
https://ci.apache.org/projects/flink/flink-docs-stable/dev/stream/state/broadcast_state.html
Am Fr., 17. Jan. 2020 um 14:50 Uhr schrieb Soheil Pourbafrani <
soheil.i...@gmail.com>:
> Hi,
>
> According to the processing logic,
Hi everyone,
The registration for Flink Forward SF 2020 is open now!
Flink Forward San Francisco 2020 will take place from March 23rd to 25th.
The conference will start with one day of training and continue with two
days of keynotes and talks.
We would like to invite you to join the Apache Flink
Hi,
The exception is thrown by Postgres.
I'd start investigating there what the problem is.
Maybe you need to tweak your Postgres configuration, but it might also be
that the Flink connector needs to be differently configured.
If the necessary config option is missing, it would be good to add.
H
Hi,
Which version are you using?
I can't find the error message in the current code base.
When writing data to a JDBC database, all Flink types must be correctly
matched to a JDBC type.
The problem is probably that Flink cannot match the 8th field of your Row
to a JDBC type.
What's the type of th
Hi Eleanore,
A dynamic filter like the one you need, is essentially a join operation.
There is two ways to do this:
* partitioning the key set and the message on the attribute. This would be
done with a KeyedCoProcessFunction.
* broadcasting the key set and just locally forwarding the messages. T
Hi,
Large state is mainly an issue for Flink's fault tolerance mechanism which
is based on periodic checkpoints, which means that the state is copied to a
remote storage system in regular intervals.
In case of a failure, the state copy needs to be loaded which takes more
time with growing state si
Hi everyone,
We know some of you only came back from holidays last week.
To give you more time to submit a talk, we decided to extend the Call for
Presentations for Flink Forward San Francisco 2020 until Sunday January
19th.
The conference takes place on March 23-25 with two days of talks and one
Hi all,
First of all, Happy New Year to everyone!
Many of you probably didn't spent the holidays thinking a lot about Flink.
Now, however, is the right time to focus again and decide which talk(s) to
submit for Flink Forward San Francisco because the Call for Presentations
is closing this Sunday,
Congrats Zhu Zhu and welcome on board!
Best, Fabian
Am Fr., 13. Dez. 2019 um 17:51 Uhr schrieb Till Rohrmann <
trohrm...@apache.org>:
> Hi everyone,
>
> I'm very happy to announce that Zhu Zhu accepted the offer of the Flink PMC
> to become a committer of the Flink project.
>
> Zhu Zhu has been
you suggested,
> https://issues.apache.org/jira/browse/FLINK-15112
>
> Many thanks,
> Chris
>
>
> ------ Original Message --
> From: "Fabian Hueske"
> To: "Chris Miller"
> Cc: "user@flink.apache.org"
> Sent: 06/12/2019 14:52:16
> S
Hi Chris,
Your query looks OK to me.
Moreover, you should get a SQLParseException (or something similar) if it
wouldn't be valid SQL.
Hence, I assume you are running in a bug in one of the optimizer rules.
I tried to reproduce the problem on the SQL training environment and
couldn't write a query
Hi,
The inline lambda MapFunction produces a Row with 12 String fields (12
calls to String.join()).
You use RowTypeInfo rowTypeDNS to declare the return type of the lambda
MapFunction. However, rowTypeDNS is defined with much more String fields.
The exception tells you that the number of fields r
Hi all,
Flink Forward North America returns to San Francisco on March 23-25, 2020.
For the first time in North America, the conference will feature two days
of talks and one day of training.
We are happy to announce that the Call for Presentations is open!
If you'd like to give a talk and share
tream with various filters applied to it. I usually see
> around 6-7 of my datastreams successfully list the JSON file in my S3
> bucket upon cancelling my Flink job.
>
>
>
> Even in my situation, would this still be an issue with S3’s file listing
> command?
>
>
>
&g
Hi,
Dynamic tables might not be persisted at all but only when it is necessary
for the computation of a query.
For example a simple "SELECT * FROM t WHERE a = 1" query on an append only
table t does not require to persist t.
However, there are a bunch of operations that require to store some part
Hi Wojciech,
I posted an answer on StackOverflow.
Best, Fabian
Am Do., 24. Okt. 2019 um 13:03 Uhr schrieb Wojciech Indyk <
wojciechin...@gmail.com>:
> Hi!
> I use Flink 1.8.0 with Kafka 2.2.1. I need to guarantee of correct order
> of events by event timestamp. I generate periodic watermarks ev
Hi Jakub,
I had a look at the changes of Flink 1.5 [1] and didn't find anything
obvious.
Something that might cause a different behavior is the new deployment and
process model (FLIP-6).
In Flink 1.5, there is a switch to disable it and use the previous
deployment mechanism.
You could try to disa
Hi Vinay,
Maybe Gordon (in CC) has an idea about this issue.
Best, Fabian
Am Do., 24. Okt. 2019 um 14:50 Uhr schrieb Vinay Patil <
vinay18.pa...@gmail.com>:
> Hi,
>
> Can someone pls help here , facing issues in Prod . I see the following
> ticket in unresolved state.
>
> https://issues.apache.
Hi,
I did not understand what you are trying to achieve.
Which field of the input table do you want to write to the output table?
Flink SQL> insert into nestedSink select nested from nestedJsonStream;
[INFO] Submitting SQL update statement to the cluster...
[ERROR] Could not execute SQL statement
Hi Fanbin,
One approach would be to ingest the field as a VARCHAR / String and
implement a Scalar UDF to convert it into a nested tuple.
The UDF could use the code of the flink-json module.
AFAIK, there is some work on the way to add built-in JSON functions.
Best, Fabian
Am Do., 24. Okt. 2019 u
Hi Komal,
Measuring latency is always a challenge. The problem here is that your
functions are chained, meaning that the result of a function is directly
passed on to the next function and only when the last function emits the
result, the first function is called with a new record.
This makes meas
Hi Michael,
One reason might be that S3's file listing command is only eventually
consistent.
It might take some time until the file appears and is listed.
Best, Fabian
Am Mi., 23. Okt. 2019 um 22:41 Uhr schrieb Nguyen, Michael <
michael.nguye...@t-mobile.com>:
> Hello all,
>
>
>
> I am running
Hi,
the exception says: "Rowtime attributes must not be in the input rows of a
regular join. As a workaround you can cast the time attributes of input
tables to TIMESTAMP before.".
The problem is that your query first joins the two tables without a
temporal condition and then wants to do a window
Hi Vishwas,
First of all, 8 GB for 60 cores is not a lot.
You might not be able to utilize all cores when running Flink.
However, the memory usage depends on several things.
Assuming your are using Flink for stream processing, the type of the state
backend is important. If you use the FSStateBack
Hi Bruce,
I haven't seen such an exception yet, but maybe Till (in CC) can help.
Best,
Fabian
Am Di., 1. Okt. 2019 um 05:51 Uhr schrieb Hanson, Bruce <
bruce.han...@here.com>:
> Hi all,
>
>
>
> We are running some of our Flink jobs with Job Manager High Availability.
> Occasionally we get a clu
Hi Oliwer,
I think you are right. There seems to be something going wrong.
Just to clarify, you are sure that the growing state size is caused by the
window operator?
>From your description I assume that the state size does not depend (solely)
on the number of distinct keys.
Otherwise, the state
Hi,
State is always associated with a single task in Flink.
The state of a task cannot be accessed by other tasks of the same operator
or tasks of other operators.
This is true for every type of state, including broadcast state.
Best, Fabian
Am Di., 1. Okt. 2019 um 08:22 Uhr schrieb Navneeth Kr
Hi,
I don' think that the memory configuration is the issue.
The problem is the join query. The join does not have any temporal
boundaries.
Therefore, both tables are completely stored in memory and never released.
You can configure a memory eviction strategy via idle state retention [1]
but you
Hi,
You enabled incremental checkpoints.
This means that parts of older checkpoints that did not change since the
last checkpoint are not removed because they are still referenced by the
incremental checkpoints.
Flink will automatically remove them once they are not needed anymore.
Are you sure t
Hi Nishant,
To answer your questions:
1) yes, the SQL time-windowed join and the DataStream API Interval Join are
the same (with different implementations though)
2) DataStream Session-window joins are not directly supported in SQL. You
can play some tricks to make it work, but it wouldn't be eleg
atch the column, so it can read the fields.
>
> Thanks for your Help!
>
> Qi Shu
>
>
> 在 2019年9月24日,下午4:36,Fabian Hueske 写道:
>
> Hi QiShu,
>
> It might be that Flink's OrcInputFormat has a bug.
> Can you open a Jira issue to report the problem?
> In order
Hi,
It's not possible to create a temporal table function from SQL, but you can
define it in the config.yaml of the SQL client as described in the
documentation [1].
Best, Fabian
[1]
https://ci.apache.org/projects/flink/flink-docs-release-1.9/dev/table/sqlClient.html#temporal-tables
Am Di., 24.
Hi,
AFAIK, Flink SQL Temporal table function joins are only supported as inner
equality joins.
An extension to left outer joins would be great, but is not on the
immediate roadmap AFAIK.
If you need the inverse, I'd recommend to implement the logic in a
DataStream program with a KeyedCoProcessFun
Hi QiShu,
It might be that Flink's OrcInputFormat has a bug.
Can you open a Jira issue to report the problem?
In order to be able to fix this, we need as much information as possible.
It would be great if you could create a minimal example of an ORC file and
a program that reproduces the issue.
If
Hi,
To expand on Dian's answer.
You should not add Flink's core libraries (APIs, core, runtime, etc.) to
your fat JAR. However, connector dependencies (like Kafka, Cassandra, etc.)
should be added.
If all your jobs require the same dependencies, you can also add JAR files
to the ./lib folder of y
rty(ProducerConfig.RETRIES_CONFIG, "3");
>
> ObjectMapper mapper = new ObjectMapper();
> DataStream sinkStreamMaliciousData = outStreamMalicious
> .map(new MapFunction,String>() {
> private static final long serialVersionUID = -6347120202L;
> @Override
> public St
Btw. there is a set difference or minus operator in the Table API [1] that
might be helpful.
[1]
https://ci.apache.org/projects/flink/flink-docs-release-1.9/dev/table/tableApi.html#set-operations
Am Fr., 20. Sept. 2019 um 15:30 Uhr schrieb Fabian Hueske :
> Hi Juan,
>
> Both,
Hi,
It depends.
There are many things that can be changed. A savepoint in Flink contains
only the state of the application and not the configuration of the system.
So an application can be migrated to another cluster that runs with a
different configuration.
There are some exceptions like the con
at least 64 bytes.
> If we have 200,000,000 per day and the allowed lateness is
> set to 7 days:
> 200,000,000 * 64 * 7 = ~83GB
>
> *For the scenario above the window metadata is useless*.
> Is there a possibility to *keep using window API*, *set allowed lateness*
> and *not
Hi Jun,
Thank you very much for your contribution.
I think a Bucketing File System Table Sink would be a great addition.
Our code contribution guidelines [1] recommend to discuss the design with
the community before opening a PR.
First of all, this ensures that the design is aligned with Flink's
Hi Juan,
Both, the local execution environment and the remote execution environment
run the same code to execute the program.
The implementation of the sortPartition operator was designed to scale to
data sizes that exceed the memory.
Internally, it serializes all records into byte arrays and sort
Hi Ken,
Changing the parallelism can affect the generation of input splits.
I had a look at BinaryInputFormat, and it adds a bunch of empty input
splits if the number of generated splits is less than the minimum number of
splits (which is equal to the parallelism).
See -->
https://github.com/apac
But with that 60 gb memory getting run out
>
> So i used below query.
> Can u please guide me in this regard
>
> On Wed, 18 Sep 2019 at 5:53 PM, Fabian Hueske wrote:
>
>> Hi,
>>
>> The query that you wrote is not a time-windowed join.
>>
>> INSERT IN
Hi,
The query that you wrote is not a time-windowed join.
INSERT INTO sourceKafkaMalicious
SELECT sourceKafka.* FROM sourceKafka INNER JOIN badips ON
sourceKafka.`source.ip`=badips.ip
WHERE sourceKafka.timestamp_received BETWEEN CURRENT_TIMESTAMP - INTERVAL
'15' MINUTE AND CURRENT_TIMESTAMP;
The
Hi,
This can be set via the environment file.
Please have a look at the documentation [1] (see "execution:
min-idle-state-retention: " and "execution: max-idle-retention: " keys).
Fabian
[1]
https://ci.apache.org/projects/flink/flink-docs-release-1.9/dev/table/sqlClient.html#environment-files
A
Hi,
No, this is not possible at the moment. You can only pass a single
expression as primary key.
A work around might be to put the two fields in a nested field (haven't
tried if this works) or combine them in a single attribute, for example by
casting them to VARCHAR and concating them.
Best, Fa
Hi,
A GROUP BY query on a streaming table requires that the result is
continuously updated.
Updates are propagated as a retraction stream (see
tEnv.toRetractStream(table, Row.class).print(); in your code).
A retraction stream encodes the type of the update as a boolean flag, the
"true" and "false
Thanks for reporting back Catlyn!
Am Do., 12. Sept. 2019 um 19:40 Uhr schrieb Catlyn Kong :
> Turns out there was some other deserialization problem unrelated to this.
>
> On Mon, Sep 9, 2019 at 11:15 AM Catlyn Kong wrote:
>
>> Hi fellow streamers,
>>
>> I'm trying to support avro BYTES type in
Congrats Zili Chen :-)
Cheers, Fabian
Am Mi., 11. Sept. 2019 um 12:48 Uhr schrieb Biao Liu :
> Congrats Zili!
>
> Thanks,
> Biao /'bɪ.aʊ/
>
>
>
> On Wed, 11 Sep 2019 at 18:43, Oytun Tez wrote:
>
>> Congratulations!
>>
>> ---
>> Oytun Tez
>>
>> *M O T A W O R D*
>> The World's Fastest Human Tran
Hi,
There is no upper limit for state size in Flink. There are applications
with 10+ TB state.
However, it is natural that checkpointing time increases with state size as
more data needs to be serialized (in case of FSStateBackend) and written to
stable storage.
(The same is btw true for recovery
Hi,
This is clearly a Scala version issue.
You need to make sure that all Flink dependencies have the same version and
are compiled for Scala 2.11.
The "_2.11" postfix in the dependency name indicates that it is a Scala
2.11 dependency ("_2.12 indicates Scala 2.12 compatibility).
Best, Fabian
Am
Hi Theo,
I would implement this with a KeyedProcessFunction.
These are the important points to consider:
1) partition the output of the Kafka source by Kafka partition (or the
attribute that determines the partition). This will ensure that the data
stay in order (per partition).
2) The KeyedProce
Hi,
that would be regular SQL cast syntax:
SELECT a, b, c, CAST(eventTime AS TIMESTAMP) FROM ...
Am Di., 10. Sept. 2019 um 18:07 Uhr schrieb Niels Basjes :
> Hi.
>
> Can you give me an example of the actual syntax of such a cast?
>
> On Tue, 10 Sep 2019, 16:30 Fabian Hueske,
database
systems have to deal with.
Best,
Fabian
Am Do., 5. Sept. 2019 um 13:37 Uhr schrieb Hanan Yehudai <
hanan.yehu...@radcom.com>:
> Thanks Fabian.
>
>
> is there any advantage using broadcast state VS using just CoMap function
> on 2 connected streams ?
>
>
>
&g
Hi Niels,
I think (not 100% sure) you could also cast the event time attribute to
TIMESTAMP before you emit the table.
This should remove the event time property (and thereby the
TimeIndicatorTypeInfo) and you wouldn't know to fiddle with the output
types.
Best, Fabian
Am Mi., 21. Aug. 2019 um 1
Hi,
CheckpointableInputFormat is only relevant if you plan to use the
InputFormat in a MonitoringFileSource, i.e., in a streaming application.
If you plan to use it in a DataSet (batch) program, InputFormat is fine.
Btw. the latest release Flink 1.9.0 has major improvements for the recovery
of ba
String key = iterator.next();
> row.setField(pos, jsonNode.get(key).asText());
> pos++;
> }
> return row;
> }
> }).returns(convert);
>
> Table tableA = tEnv.fromDataStream(dataStreamRow);
>
>
> Le jeu. 5 sept. 2019 à 13:23, Fabian Hueske a écr
Hi everyone,
I'm very happy to announce that Kostas Kloudas is joining the Flink PMC.
Kostas is contributing to Flink for many years and puts lots of effort in
helping our users and growing the Flink community.
Please join me in congratulating Kostas!
Cheers,
Fabian
Hi,
Kostas (in CC) might be able to help.
Best, Fabian
Am Mi., 4. Sept. 2019 um 22:59 Uhr schrieb sidhartha saurav <
sidsau...@gmail.com>:
> Hi,
>
> Can someone suggest a workaround so that we do not get this issue while
> changing the S3 bucket ?
>
> On Thu, Aug 22, 2019 at 4:24 PM sidhartha s
Hi Steve,
Maybe you could implement a custom TableSource that queries the data from
the rest API and converts the JSON directly into a Row data type.
This would also avoid going through the DataStream API just for ingesting
the data.
Best, Fabian
Am Mi., 4. Sept. 2019 um 15:57 Uhr schrieb Steve
Hi,
Are you getting this error repeatedly or was this a single time?
If it's just a single time error, it's probably caused by a task manager
process that died for some reason (as suggested by the error message).
You should have a look at the TM logs whether you can finds something that
would exp
Hi Vishwas,
This is a log statement from Kafka [1].
Not sure how when AppInfoParser is created (the log message is written by
the constructor).
For Kafka versions > 1.0, I'd recommend the universal connector [2].
Not sure how well it works if producers and consumers have different
versions.
Mayb
Hi,
A window needs to keep the data as long as it expects new data.
This is clearly the case before the end time of the window was reached. If
my window ends at 12:30, I want to wait (at least) until 12:30 before I
remove any data, right?
In case you expect some data to be late, you can configure
Hi,
Flink does not have good support for mixing bounded and unbounded streams
in its DataStream API yet.
If the dimension table is static (and small enough), I'd use a
RichMapFunction and load the table in the open() method into the heap.
In this case, you'd probably need to restart the job (can b
Hi all,
The registration for the Flink Forward Europe training sessions closes in
four weeks.
The training takes place in Berlin at October 7th and is followed by two
days of talks by speakers from companies like Airbus, Goldman Sachs,
Netflix, Pinterest, and Workday [1].
The following four train
9 Uhr schrieb Hanan Yehudai <
hanan.yehu...@radcom.com>:
> Im not sure what you mean by use process function and not window process
> function , as the window operator takes in a windowprocess function..
>
>
>
> *From:* Fabian Hueske
> *Sent:* Monday, August 26, 20
up with a partition
>> containing element out of ‘window order’.
>>
>> I was also thinking this problem is very similar to that of checkpoint
>> barriers. I intended to dig into the details of the exactly once Kafka sink
>> for some inspiration.
>>
>> Padarn
>
Hi Sushant,
It's hard to tell what's going on.
Maybe the thread pool of the async io operator is too small for the
ingested data rate?
This could cause the backpressure on the source and eventually also the
failing checkpoints.
Which Flink version are you using?
Best, Fabian
Am Do., 29. Aug. 2
D*
> The World's Fastest Human Translation Platform.
> oy...@motaword.com — www.motaword.com
>
>
> On Tue, Aug 27, 2019 at 8:11 AM Fabian Hueske wrote:
>
>> Hi all,
>>
>> Flink 1.9 Docker images are available at Docker Hub [1] now.
>> Due to some configur
eems to have some quirks).
>
> I think ideally each partition of the kafka topic would have some regular
> information about watermarks. Perhaps the kafka producer can be modified to
> support this.
>
> Padarn
>
> On Fri, Aug 16, 2019 at 3:50 PM Fabian Hueske wrote:
>
>&
27;t output a result any more when testing all of those
> combinations. Now the second attempt works but isn't really what I wanted
> to query (as the "same day"-predicate is still missing).
>
> Best regards
> Theo
>
> --
> *Von: *&qu
Hi Theo,
The work on custom triggers has been put on hold due to some major
refactorings (splitting the modules, porting Scala code to Java, new type
system, new catalog interfaces, integration of the Blink planner).
It's also not on the near-time roadmap AFAIK.
To be honest, I'm not sure how much
Hi all,
Flink 1.9 Docker images are available at Docker Hub [1] now.
Due to some configuration issue, there are only Scala 2.11 issues at the
moment but this was fixed [2].
Flink 1.9 Scala 2.12 images should be available soon.
Cheers,
Fabian
[1] https://hub.docker.com/_/flink
[2]
https://github.
ng on Flink’s monitoring page - for the watermarks I see
> different vales even after all my files were processed. Which is
> something I would not expect
> I would expect that eventually the WM will be the highest EVENT_TIME on
> my set of files..
>
>
>
>
>
> than
1 - 100 of 1728 matches
Mail list logo