Google Dataflow currently uses a JSON representation of the pipeline graph and also the pipeline proto. We represent the graph in two different ways which leads to some wonderful *features*. Google Dataflow also side steps the Beam job service since Dataflow has its own Job API. Supporting the Beam Job API is planned future work though since we would like to use its representation but the Beam Job API lacks APIs that Dataflow already provides to users via its own Job API.
On the pipeline execution side, Google Dataflow uses containers exclusively for Google Cloud customers and the Fn API to process bundles, perform logging, side input lookups, timers, user state. So yes, Google Dataflow is a hybrid today but wishes to move to full portability when it can. On Wed, Sep 18, 2019 at 8:27 AM Chad Dombrova <chad...@gmail.com> wrote: > Just note that while Dataflow does have robust python support it does not > fully support the portability framework. It’s a bit of a blurry > distinction, and honestly I’m not crystal clear on this as I get the > impression that Dataflow may be a bit of a Portability hybrid. It does not > use the job service or the expansion service but I have heard that it uses > the external worker pool, and maybe it uses the protobufs pipeline > definitions? It’d be great if a Googler could clarify. > > -chad > > > On Wed, Sep 18, 2019 at 7:51 AM Holden Karau <hol...@pigscanfly.ca> wrote: > >> Probably the most stable is running on Dataflow still. But I’m excited to >> see the progress towards a Spark runner, can’t wait to try TFT on it :) >> >> On Tue, Sep 17, 2019 at 4:37 PM Kyle Weaver <kcwea...@google.com> wrote: >> >>> The Flink runner is definitely more stable, as it's been around for >>> longer and has more developers and users on it. But a lot of the code is >>> shared, so for example some of the issues above would also happen on the >>> Flink runner. >>> >>> Kyle Weaver | Software Engineer | github.com/ibzib | kcwea...@google.com >>> >>> >>> On Tue, Sep 17, 2019 at 4:34 PM Benjamin Tan < >>> benjamintanwei...@gmail.com> wrote: >>> >>>> Thanks for all the replies Kyle! You've been super helpful :D. >>>> >>>> Would you say that the Flink runner more stable than the Spark one? Or >>>> which combo is the most stable for now? >>>> >>>> On 2019/09/17 19:43:54, Tom Barber <t...@spicule.co.uk> wrote: >>>> > Thanks Kyle, >>>> > >>>> > From my pov Alpha is fine, I’m just trying to test out some of the >>>> > capabilities currently, but trying to dig around the website doesn’t >>>> > explain a great deal. Luckily Benjamin seems a step ahead of me… I >>>> hope it >>>> > stays that way! ;) >>>> > >>>> > >>>> > On 17 September 2019 at 19:33:40, Kyle Weaver (kcwea...@google.com) >>>> wrote: >>>> > >>>> > > The amount of issues I've encountered as a newbie is indeed >>>> troubling. >>>> > Spark portability is very much "alpha" quality software, a point we >>>> should >>>> > maybe emphasize on the website more. Anyway, I appreciate your >>>> patience, >>>> > and I'll do my best to address all these issues. >>>> > >>>> > > org.apache.beam.vendor.grpc.v1p21p0.io.grpc.StatusRuntimeException: >>>> > UNIMPLEMENTED: Method not found! >>>> > > AttributeError: module 'apache_beam.coders.coders' has no >>>> > attribute ‘VarIntCoder' >>>> > This class of errors occurs when the SDK version does not match up >>>> with the >>>> > runner version -- unfortunately, we cannot guarantee compatibility >>>> between >>>> > the two, so manual syncing is required for now. We are looking for >>>> ways to >>>> > improve this. >>>> > >>>> > > ERROR:grpc._server:Exception calling application: u'2-1' >>>> > It looks like these errors are entirely spurious. Have a fix for it >>>> here: >>>> > https://github.com/apache/beam/pull/9600 >>>> > >>>> > Note that there may still be other spurious errors like: >>>> > >>>> > 19/09/17 11:25:31 ERROR ManagedChannelOrphanWrapper: *~*~*~ Channel >>>> > ManagedChannelImpl{logId=84, target=localhost:36129} was not shutdown >>>> > properly!!! ~*~*~* >>>> > >>>> > and >>>> > >>>> > 19/09/17 11:25:32 ERROR SerializingExecutor: Exception while executing >>>> > runnable >>>> > >>>> org.apache.beam.vendor.grpc.v1p21p0.io.grpc.internal.ServerImpl$JumpToApplicationThreadServerStreamListener$1HalfClosed@47db89b1 >>>> > java.lang.IllegalStateException: call already closed >>>> > >>>> > I will try to get rid of them also, but for now you can just ignore >>>> them. >>>> > They are annoying but harmless. >>>> > >>>> > Kyle Weaver | Software Engineer | github.com/ibzib | >>>> kcwea...@google.com >>>> > >>>> > >>>> > On Tue, Sep 17, 2019 at 8:05 AM Benjamin Tan < >>>> benjamintanwei...@gmail.com> >>>> > wrote: >>>> > >>>> > > :D. Still, I'm curious as to the error we both are getting. Maybe >>>> someone >>>> > > can shed some light on it. >>>> > > >>>> > > On Tue, Sep 17, 2019 at 10:54 PM Tom Barber <t...@spicule.co.uk> >>>> wrote: >>>> > > >>>> > >> I do see hello written to 1 file and world to another, I guess it >>>> works! >>>> > >> Thanks for the pointers Benjamin I was about to give up. >>>> > >> >>>> > >> Tom >>>> > >> >>>> > >> >>>> > >> On 17 September 2019 at 15:51:13, Benjamin Tan ( >>>> > >> benjamintanwei...@gmail.com) wrote: >>>> > >> >>>> > >> Tell me if you see any output. Anyway, here's the link to the same >>>> issue >>>> > >> you're facing: >>>> > >> >>>> > >> >>>> > >> >>>> https://lists.apache.org/thread.html/4e8e1455916debe096de32551f9ab05853524cf282bc312cd4620d68@%3Cuser.beam.apache.org%3E >>>> > >> >>>> > >> The amount of issues I've encountered as a newbie is indeed >>>> troubling. >>>> > >> >>>> > >> On 2019/09/17 14:43:11, Tom Barber <t...@spicule.co.uk> wrote: >>>> > >> > 🤣 okay I’ll look again, I assumed it just crashed in a ball of >>>> flames! >>>> > >> > >>>> > >> > >>>> > >> > On 17 September 2019 at 15:39:33, Benjamin Tan ( >>>> > >> benjamintanwei...@gmail.com) >>>> > >> > wrote: >>>> > >> > >>>> > >> > I got this too! Did you manage to get any output? (I did) I >>>> reported >>>> > >> this >>>> > >> > in another thread. >>>> > >> > >>>> > >> > This looks like a key error when StopWorker is called. As far as >>>> I >>>> > >> know, it >>>> > >> > seems like the work has been processed. >>>> > >> > >>>> > >> > On Tue, Sep 17, 2019 at 10:31 PM Tom Barber <t...@spicule.co.uk> >>>> wrote: >>>> > >> > >>>> > >> > > Well my errors are different but still terminal: >>>> > >> > > >>>> > >> > > ERROR:grpc._server:Exception calling application: u'1-1' >>>> > >> > > Traceback (most recent call last): >>>> > >> > > File "/usr/local/lib/python2.7/dist-packages/grpc/_server.py", >>>> line >>>> > >> 434, >>>> > >> > > in _call_behavior >>>> > >> > > response_or_iterator = behavior(argument, context) >>>> > >> > > File >>>> > >> > > >>>> > >> >>>> "/usr/local/lib/python2.7/dist-packages/apache_beam-2.17.0.dev0-py2.7.egg/apache_beam/runners/worker/worker_pool_main.py", >>>> > >> > > line 126, in StopWorker >>>> > >> > > worker_process = >>>> > >> > > self._worker_processes.pop(stop_worker_request.worker_id) >>>> > >> > > KeyError: u'1-1' >>>> > >> > > ERROR:grpc._server:Exception calling application: u'2-1' >>>> > >> > > Traceback (most recent call last): >>>> > >> > > File "/usr/local/lib/python2.7/dist-packages/grpc/_server.py", >>>> line >>>> > >> 434, >>>> > >> > > in _call_behavior >>>> > >> > > response_or_iterator = behavior(argument, context) >>>> > >> > > File >>>> > >> > > >>>> > >> >>>> "/usr/local/lib/python2.7/dist-packages/apache_beam-2.17.0.dev0-py2.7.egg/apache_beam/runners/worker/worker_pool_main.py", >>>> > >> > > line 126, in StopWorker >>>> > >> > > worker_process = >>>> > >> > > self._worker_processes.pop(stop_worker_request.worker_id) >>>> > >> > > KeyError: u'2-1' >>>> > >> > > >>>> > >> > > >>>> > >> > > On 17 September 2019 at 14:46:12, Benjamin Tan ( >>>> > >> > > benjamintanwei...@gmail.com) wrote: >>>> > >> > > >>>> > >> > > If it helps, I’m using Spark 2.4.4. The Apache Beam Python >>>> library on >>>> > >> > > master is 2.17.0-dev. >>>> > >> > > >>>> > >> > > >>>> > >> > > On 17 Sep 2019, at 9:39 PM, Tom Barber <t...@spicule.co.uk> >>>> wrote: >>>> > >> > > >>>> > >> > > Cool thanks Benjamin, I’ll give it a shot. >>>> > >> > > >>>> > >> > > Tom >>>> > >> > > >>>> > >> > > >>>> > >> > > On 17 September 2019 at 13:56:14, Benjamin Tan ( >>>> > >> > > benjamintanwei...@gmail.com) wrote: >>>> > >> > > >>>> > >> > > >>>> > >> > > I encountered the exact same thing today. High five! Here’s >>>> how I >>>> > >> managed >>>> > >> > > to make some progress: >>>> > >> > > >>>> > >> > > 1. Used the master branch >>>> > >> > > 2. Built and installed the Python SDK >>>> > >> > > >>>> > >> > > cd into the sdk library and python ./setup install >>>> > >> > > >>>> > >> > > I got some other errors but they didn’t seem to be show >>>> stoppers. >>>> > >> > > >>>> > >> > > On 17 Sep 2019, at 8:26 PM, Tom Barber <t...@spicule.co.uk> >>>> wrote: >>>> > >> > > >>>> > >> > > Hello folks, >>>> > >> > > >>>> > >> > > Day 3 of trying to get the basics going with Python & Spark >>>> 2.2.3. >>>> > >> > > >>>> > >> > > I’ve downgraded the spark version to 2.2.3 in the cradle build >>>> so >>>> > >> that I >>>> > >> > > can run jobs against it. >>>> > >> > > >>>> > >> > > I’ve then written this: >>>> > >> > > >>>> > >> > > options = PipelineOptions(["--runner=PortableRunner", >>>> > >> > > "--job_endpoint=localhost:8099", >>>> "--environment_type=LOOPBACK"]) >>>> > >> > > >>>> > >> > > with beam.Pipeline(options=options) as p: >>>> > >> > > print("hello") >>>> > >> > > lines = p | 'Create' >> beam.Create(['hello', 'world']) >>>> > >> > > >>>> > >> > > Which in turn explodes with: >>>> > >> > > >>>> > >> > > Caused by: >>>> > >> > > >>>> org.apache.beam.vendor.grpc.v1p21p0.io.grpc.StatusRuntimeException: >>>> > >> > > UNIMPLEMENTED: Method not found! >>>> > >> > > at >>>> > >> > > >>>> > >> >>>> org.apache.beam.vendor.grpc.v1p21p0.io.grpc.stub.ClientCalls.toStatusRuntimeException(ClientCalls.java:235) >>>> > >> > > at >>>> > >> > > >>>> > >> >>>> org.apache.beam.vendor.grpc.v1p21p0.io.grpc.stub.ClientCalls.getUnchecked(ClientCalls.java:216) >>>> > >> > > at >>>> > >> > > >>>> > >> >>>> org.apache.beam.vendor.grpc.v1p21p0.io.grpc.stub.ClientCalls.blockingUnaryCall(ClientCalls.java:141) >>>> > >> > > at >>>> > >> > > >>>> > >> >>>> org.apache.beam.model.fnexecution.v1.BeamFnExternalWorkerPoolGrpc$BeamFnExternalWorkerPoolBlockingStub.startWorker(BeamFnExternalWorkerPoolGrpc.java:226) >>>> > >> > > >>>> > >> > > When run against python 2 and >>>> > >> > > >>>> > >> > > AttributeError: module 'apache_beam.coders.coders' has no >>>> > >> > > attribute ‘VarIntCoder' >>>> > >> > > >>>> > >> > > When run with python 3.5. >>>> > >> > > >>>> > >> > > Should I just give up trying the portable runner at the moment >>>> or am I >>>> > >> > > doing something dumb/incompatible? I’m stuck on Spark 2.2.3 so >>>> > >> upgrading >>>> > >> > > isn’t currently an option. >>>> > >> > > >>>> > >> > > Tom >>>> > >> > > >>>> > >> > > Spicule Limited is registered in England & Wales. Company >>>> Number: >>>> > >> > > 09954122. Registered office: First Floor, Telecom House, >>>> 125-135 >>>> <https://www.google.com/maps/search/125-135+%0D%0A+Preston+%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g> >>>> >>>> <https://www.google.com/maps/search/125-135+%0D%0A+Preston+%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g>> >>>> >> Preston >>>> <https://www.google.com/maps/search/125-135+%0D%0A+Preston+%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g> >>>> >>>> <https://www.google.com/maps/search/125-135+%0D%0A+Preston+%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g>> >>>> >> > > Road, Brighton, England, BN1 6AF >>>> <https://www.google.com/maps/search/125-135+%0D%0A+Preston+%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g>. >>>> VAT No. 251478891. >>>> > >> > > >>>> > >> > > >>>> > >> > > All engagements are subject to Spicule Terms and Conditions of >>>> > >> Business. >>>> > >> > > This email and its contents are intended solely for the >>>> individual to >>>> > >> whom >>>> > >> > > it is addressed and may contain information that is >>>> confidential, >>>> > >> > > privileged or otherwise protected from disclosure, >>>> distributing or >>>> > >> copying. >>>> > >> > > Any views or opinions presented in this email are solely those >>>> of the >>>> > >> > > author and do not necessarily represent those of Spicule >>>> Limited. The >>>> > >> > > company accepts no liability for any damage caused by any virus >>>> > >> transmitted >>>> > >> > > by this email. If you have received this message in error, >>>> please >>>> > >> notify us >>>> > >> > > immediately by reply email before deleting it from your system. >>>> > >> Service of >>>> > >> > > legal notice cannot be effected on Spicule Limited by email. >>>> > >> > > >>>> > >> > > >>>> > >> > > Spicule Limited is registered in England & Wales. Company >>>> Number: >>>> > >> > > 09954122. Registered office: First Floor, Telecom House, >>>> 125-135 >>>> <https://www.google.com/maps/search/125-135+%0D%0A+Preston+%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g> >>>> >>>> <https://www.google.com/maps/search/125-135+%0D%0A+Preston+%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g>> >>>> >> Preston >>>> <https://www.google.com/maps/search/125-135+%0D%0A+Preston+%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g> >>>> >>>> <https://www.google.com/maps/search/125-135+%0D%0A+Preston+%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g>> >>>> >> > > Road, Brighton, England, BN1 6AF >>>> <https://www.google.com/maps/search/125-135+%0D%0A+Preston+%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g>. >>>> VAT No. 251478891. >>>> > >> > > >>>> > >> > > >>>> > >> > > All engagements are subject to Spicule Terms and Conditions of >>>> > >> Business. >>>> > >> > > This email and its contents are intended solely for the >>>> individual to >>>> > >> whom >>>> > >> > > it is addressed and may contain information that is >>>> confidential, >>>> > >> > > privileged or otherwise protected from disclosure, >>>> distributing or >>>> > >> copying. >>>> > >> > > Any views or opinions presented in this email are solely those >>>> of the >>>> > >> > > author and do not necessarily represent those of Spicule >>>> Limited. The >>>> > >> > > company accepts no liability for any damage caused by any virus >>>> > >> transmitted >>>> > >> > > by this email. If you have received this message in error, >>>> please >>>> > >> notify us >>>> > >> > > immediately by reply email before deleting it from your system. >>>> > >> Service of >>>> > >> > > legal notice cannot be effected on Spicule Limited by email. >>>> > >> > > >>>> > >> > > >>>> > >> > > Spicule Limited is registered in England & Wales. Company >>>> Number: >>>> > >> > > 09954122. Registered office: First Floor, Telecom House, >>>> 125-135 >>>> <https://www.google.com/maps/search/125-135+%0D%0A+Preston+%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g> >>>> >>>> <https://www.google.com/maps/search/125-135+%0D%0A+Preston+%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g>> >>>> >> Preston >>>> <https://www.google.com/maps/search/125-135+%0D%0A+Preston+%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g> >>>> >>>> <https://www.google.com/maps/search/125-135+%0D%0A+Preston+%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g>> >>>> >> > > Road, Brighton, England, BN1 6AF >>>> <https://www.google.com/maps/search/125-135+%0D%0A+Preston+%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g>. >>>> VAT No. 251478891. >>>> > >> > > >>>> > >> > > >>>> > >> > > All engagements are subject to Spicule Terms and Conditions of >>>> > >> Business. >>>> > >> > > This email and its contents are intended solely for the >>>> individual to >>>> > >> whom >>>> > >> > > it is addressed and may contain information that is >>>> confidential, >>>> > >> > > privileged or otherwise protected from disclosure, >>>> distributing or >>>> > >> copying. >>>> > >> > > Any views or opinions presented in this email are solely those >>>> of the >>>> > >> > > author and do not necessarily represent those of Spicule >>>> Limited. The >>>> > >> > > company accepts no liability for any damage caused by any virus >>>> > >> transmitted >>>> > >> > > by this email. If you have received this message in error, >>>> please >>>> > >> notify us >>>> > >> > > immediately by reply email before deleting it from your system. >>>> > >> Service of >>>> > >> > > legal notice cannot be effected on Spicule Limited by email. >>>> > >> > > >>>> > >> > >>>> > >> > -- >>>> > >> > >>>> > >> > >>>> > >> > Spicule Limited is registered in England & Wales. Company Number: >>>> > >> > 09954122. Registered office: First Floor, Telecom House, 125-135 >>>> Preston >>>> <https://www.google.com/maps/search/125-135+Preston+%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g> >>>> >>>> <https://www.google.com/maps/search/125-135+Preston+%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g>> >>>> >> > Road, Brighton, England, BN1 6AF >>>> <https://www.google.com/maps/search/125-135+Preston+%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g>. >>>> VAT No. 251478891. >>>> > >> > >>>> > >> > >>>> > >> > >>>> > >> > >>>> > >> > All engagements >>>> > >> > are subject to Spicule Terms and Conditions of Business. This >>>> email and >>>> > >> its >>>> > >> > contents are intended solely for the individual to whom it is >>>> addressed >>>> > >> and >>>> > >> > may contain information that is confidential, privileged or >>>> otherwise >>>> > >> > protected from disclosure, distributing or copying. Any views or >>>> > >> opinions >>>> > >> > presented in this email are solely those of the author and do not >>>> > >> > necessarily represent those of Spicule Limited. The company >>>> accepts no >>>> > >> > liability for any damage caused by any virus transmitted by this >>>> email. >>>> > >> If >>>> > >> > you have received this message in error, please notify us >>>> immediately by >>>> > >> > reply email before deleting it from your system. Service of >>>> legal notice >>>> > >> > cannot be effected on Spicule Limited by email. >>>> > >> > >>>> > >> >>>> > >> >>>> > >> Spicule Limited is registered in England & Wales. Company Number: >>>> > >> 09954122. Registered office: First Floor, Telecom House, 125-135 >>>> Preston >>>> <https://www.google.com/maps/search/125-135+Preston+%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g> >>>> >>>> <https://www.google.com/maps/search/125-135+Preston+%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g>> >>>> >> Road, Brighton, England, BN1 6AF >>>> <https://www.google.com/maps/search/125-135+Preston+%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g>. >>>> VAT No. 251478891. >>>> > >> >>>> > >> >>>> > >> All engagements are subject to Spicule Terms and Conditions of >>>> Business. >>>> > >> This email and its contents are intended solely for the individual >>>> to whom >>>> > >> it is addressed and may contain information that is confidential, >>>> > >> privileged or otherwise protected from disclosure, distributing or >>>> copying. >>>> > >> Any views or opinions presented in this email are solely those of >>>> the >>>> > >> author and do not necessarily represent those of Spicule Limited. >>>> The >>>> > >> company accepts no liability for any damage caused by any virus >>>> transmitted >>>> > >> by this email. If you have received this message in error, please >>>> notify us >>>> > >> immediately by reply email before deleting it from your system. >>>> Service of >>>> > >> legal notice cannot be effected on Spicule Limited by email. >>>> > >> >>>> > > >>>> > >>>> > -- >>>> > >>>> > >>>> > Spicule Limited is registered in England & Wales. Company Number: >>>> > 09954122. Registered office: First Floor, Telecom House, 125-135 >>>> Preston >>>> <https://www.google.com/maps/search/125-135+Preston++%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g> >>>> >>>> <https://www.google.com/maps/search/125-135+Preston++%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g>> >>>> Road, Brighton, England, BN1 6AF >>>> <https://www.google.com/maps/search/125-135+Preston++%0D%0A+Road,+Brighton,+England,+BN1+6AF?entry=gmail&source=g>. >>>> VAT No. 251478891. >>>> > >>>> > >>>> > >>>> > >>>> > All engagements >>>> > are subject to Spicule Terms and Conditions of Business. This email >>>> and its >>>> > contents are intended solely for the individual to whom it is >>>> addressed and >>>> > may contain information that is confidential, privileged or otherwise >>>> > protected from disclosure, distributing or copying. Any views or >>>> opinions >>>> > presented in this email are solely those of the author and do not >>>> > necessarily represent those of Spicule Limited. The company accepts >>>> no >>>> > liability for any damage caused by any virus transmitted by this >>>> email. If >>>> > you have received this message in error, please notify us immediately >>>> by >>>> > reply email before deleting it from your system. Service of legal >>>> notice >>>> > cannot be effected on Spicule Limited by email. >>>> > >>>> >>> -- >> Twitter: https://twitter.com/holdenkarau >> Books (Learning Spark, High Performance Spark, etc.): >> https://amzn.to/2MaRAG9 <https://amzn.to/2MaRAG9> >> YouTube Live Streams: https://www.youtube.com/user/holdenkarau >> >