Of course, 15 minutes after I give up and decide to email the mailing list,
I figure it out - my flink App was using the CollectionsEnvironment and not
the proper RemoteEnvironment.
It is still the case, however, that the `flink-table` JAR built by the
standard commands doesn't include the
Hello!
We are attempting to use the Flink Table API, but are running into a few
issues.
We initially started with our dependencies looking something like this:
libraryDependencies ++= Seq(
"org.apache.flink" %% "flink-scala" % "1.2.0" % "provided",
"org.apache.flink" %% "flink-clients" %
Hi Ethan,
I believe then it is because the Watermark and Timestamps in your
implementation are uncorrelated. What Watermark really is a marker that
says there will be no elements with timestamp smaller than the value of
this watermark. For more info on the concept see [1]
Hi all,
I would appreciate for any help or advice in regard to default Java runtime
shutdown hooks and canceling Flink jobs.
Namely part of my Flink application I am using a Kafka interceptor class that
defines a shutdown hook thread. When stopping the Flink streaming job on my
local machine
Hi Dawid, I'm working with Max on the project
Our code for the TimestampAndWatermarkAssigner is:
```
class TimestampAndWatermarkAssigner(val maxLateness: Long) extends
AssignerWithPeriodicWatermarks[Row] {
override def extractTimestamp(element: Row, previousElementTimestamp:
Long): Long = {
Thanks for your reply. It makes things much clearer for me. I think you are
right - Side Inputs are probably the right way long term (I looked at the Team
definition), but I think I can construct something in the mean time.
Steve
On Mar 7, 2017, at 6:11 AM, Aljoscha Krettek
Hi Max,
How do you assign timestamps to your events (in event-time case)? Could you
post whole code for your TimestampAndWatermarkAssigner?
Regards,
Dawid
2017-03-07 20:59 GMT+01:00 ext.mwalker :
> Hi Stephan,
>
> The right number of events seem to leave the source
Hi Stephan,
The right number of events seem to leave the source and enter the windows,
but it shows that 0 exit the windows.
Also I have tried 30 minutes and not setting the watermark interval, I am
not sure what I am supposed to put there the docs seem vague about that.
Best,
Max
On Tue, Mar
Hi Sathi,
if you read data from Kinesis than Flink can offer you exactly once
processing guarantees. However, what you see written out to your database
depends a little bit on the implementation of your custom sink. If you have
synchronous JDBC client which does not lose data and you fail your
> is there some shading logic involved in the dependencies, concerning the AWS
> libraries?
Not that I am aware of. The AWS code is included in the job's fat jar as-is.
Hi!
At a first glance, your code looks correct to assign the Watermarks. What
is your watermark interval in the config?
Can you check with the Flink metrics (if you are using 1.2) to see how many
rows leave the source, how many enter/leave the window operators, etc?
That should help figuring
Hi all,
I want to label events in a stream based on a condition on some future
events.
For example my stream contains events of type A and B and and I would like
to assign a label 1 to an event E of type A if an event of type B happens
within a duration x of E. I am using event time and my events
I'll try and add more details in a bit.
If you have some suggestions on how to make the serialization stack more
extensible, please let us know!
Some hooks exist, like TypeInfoFactories:
Great to hear it!
What do you think about adding a section to the Flink docs about deployment
on Azure (there is already AWS and GCE, so Azure would make the
cloud-trinity complete) that explains how to set this up and avoid such
pitfalls.
@Shannon @Gordon - is there some shading logic involved in the
dependencies, concerning the AWS libraries?
On Tue, Mar 7, 2017 at 5:50 PM, Tzu-Li (Gordon) Tai
wrote:
> Hi,
>
> I just had a quick look on this, but the Kafka fetcher thread’s context
> classloader doesn’t
Generally, using S3 filesystem in EMR with Flink has worked pretty well for me
in Flink < 1.2 (unless you run out of connections in your HTTP pool). When you
say, "using Hadoop File System class", what do you mean? In my experience, it's
sufficient to just use the "s3://" filesystem protocol
Hi Gordon,
Thanks for the advice. Following it I’ve implemented the
Keyed(De)SerializationSchema and am able to further emit the metadata to
downstream operators.
Regards,
Dominik
> On 7 Mar 2017, at 07:08, Tzu-Li (Gordon) Tai wrote:
>
> Hi Dominik,
>
> I would
Hi,
I just had a quick look on this, but the Kafka fetcher thread’s context
classloader doesn’t seem to be the issue (at least for 1.1.4).
In Flink 1.1.4, a separate thread from the task thread is created to run the
fetcher, but since the task thread sets the user code classloader as its
Hi Gordon,
Many thanks for your helpful ideas. We tried yesterday the CEP approach,
but could not figure it out. The ProcessFunction one looks more promising,
and we are investigating it, though we are fighting with some issues
related to the event time, where we cannot see so far the timer
Hi Guys,
Has anyone got this error before ? If yes, have you found any other
solution apart from copying the jar files to flink lib folder
Regards,
Vinay Patil
On Mon, Mar 6, 2017 at 8:21 PM, vinay patil [via Apache Flink User Mailing
List archive.]
Ah, I see...
The issue is that the Kafka fetcher thread apparently do not have the
user-code class loader set as the context class loader. Kryo relies on that
for class resolution.
What Flink version are you on? I think that actual processing and
forwarding does not happen in the Kafka Fetchers
21 matches
Mail list logo