Clarifying: Raghu started working on this code before we dropped the code to Beam, so we're finishing up code review there and then we'll actually commit it to the Beam repository.
As far as we know, however, the code should be completely interoperable with Beam today and it will soon be coming as a PR against apache/incubator-beam repo. Thanks! Dan On Fri, Mar 18, 2016 at 11:47 AM, Raghu Angadi <[email protected]> wrote: > Thanks for trying it. > > I fixed the CheckStyle error (not sure why my build is not failing). Let > me know if you see any issues running with Beam. I haven't tried it. I > should. In fact Daniel Halperin says my patch should be against Beam.. > > Raghu. > > On Fri, Mar 18, 2016 at 11:22 AM, William McCarthy < > [email protected]> wrote: > >> Thanks JB, Emanuele, Kostas & Raghu, especially Kostas and Raghu for >> pointing me to working code. >> >> I’m in the middle of a hack day at the moment, so the speed of your >> responses has been very welcome. >> >> In the first instance, I’ll try using your changes, Raghu. I’ve cloned >> your repo, switched to the kafka branch and built both contrib/kafka and >> contrib/examples/kafka. The contrib/kafka initially failed with a >> CheckStyle error >> (/Users/bill/dev/DataflowJavaSDK/contrib/kafka/src/main/java/com/google/cloud/dataflow/contrib/kafka/KafkaIO.java:683:12: >> 'private' modifier out of order with the JLS suggestions)… I’ve fixed that >> in my local clone and now it’s building fine. I hope to be able to run your >> contrib unchanged on top of the incubator-beam codebase, which will be what >> I attempt to do now. >> >> Thanks again to all, for your swift help. >> >> Bill >> >> On Mar 18, 2016, at 12:55 PM, Raghu Angadi <[email protected]> wrote: >> >> Hi Bill, >> >> We have fairly well tested patch for KafkaIO (pr #121 >> <https://github.com/GoogleCloudPlatform/DataflowJavaSDK/pull/121>). It >> will be merged soon. The example there keeps track of top hashtags in 10 >> minute sliding window and writes the results to another Kafka topic. Please >> try it if you can. It is well tested on Google Cloud Dataflow. I have not >> run it using Flink runner. >> >> Raghu. >> >> On Fri, Mar 18, 2016 at 9:46 AM, Kostas Kloudas < >> [email protected]> wrote: >> >>> Hello Bill, >>> >>> This is a known limitation of the Flink Runner. >>> There is a JIRA issue for that >>> https://issues.apache.org/jira/browse/BEAM-127 >>> >>> A wrapper for Flink sinks will come soon and as Beam evolves, >>> a more Beam-y solution will come as well. >>> >>> Kostas >>> >>> On Mar 18, 2016, at 5:23 PM, William McCarthy <[email protected]> >>> wrote: >>> >>> Hi, >>> >>> I’m trying to write a proof-of-concept which takes messages from Kafka, >>> transforms them using Beam on Flink, then pushes the results onto a >>> different Kafka topic. >>> >>> I’ve used the KafkaWindowedWordCountExample as a starting point, and >>> that’s doing the first part of what I want to do, but it outputs to text >>> files as opposed to Kafka. FlinkKafkaProducer08 looks promising, but I >>> can’t figure out how to plug it into the pipeline. I was thinking that it >>> would be wrapped with an UnboundedFlinkSink, or some such, but that doesn’t >>> seem to exist. >>> >>> Any advice or thoughts on what I’m trying to do? >>> >>> I’m running the latest incubator-beam (as of last night from Github), >>> Flink 1.0.0 in cluster mode and Kafka 0.9.0.1, all on Google Compute Engine >>> (Debian Jessie). >>> >>> Thanks, >>> >>> Bill McCarthy >>> >>> >>> >> >> >
