Hi Fabian,
Thank you for quick response. We will check the versions of the flink in cluster and maven. By the way, the BucketingSink function is really good at sinking to hdfs (Especially for event time file naming). However, currently in version 1.1.3, the RollingSink function is used. Is there any plan to release the BucketingSink function in stable version? ________________________________ Gönderen: Fabian Hueske <fhue...@gmail.com> Gönderildi: 26 Ekim 2016 Çarşamba 20:38:56 Kime: dev@flink.apache.org Konu: Re: Flink 1.2-SNAPSHOT Bucketing Sink problem Hi Ozan, a NoSuchMethodError indicates a version mismatch. Since you are using a SNAPSHOT build it is likely that dependencies changed when you recompiled Flink or your job. You should make sure that you are using the same version (on SNAPSHOT the same commit) for jobs and cluster. Best, Fabian 2016-10-26 18:30 GMT+02:00 Ozan DENİZ <ozande...@outlook.com>: > Hi everyone, > > > We are currently using BucketingSink (Flink 1.2-SNAPSHOT) function for > sink operation. The BucketingSink function was working correctly. However > today, when I put the job on cluster I got this error; > > > java.lang.NoSuchMethodError: org.apache.flink.streaming.api.operators. > StreamingRuntimeContext.getProcessingTimeService() > Lorg/apache/flink/streaming/runtime/tasks/ProcessingTimeService; > at org.apache.flink.streaming.connectors.fs.bucketing. > BucketingSink.open(BucketingSink.java:327) > at org.apache.flink.api.common.functions.util.FunctionUtils. > openFunction(FunctionUtils.java:36) > at org.apache.flink.streaming.api.operators. > AbstractUdfStreamOperator.open(AbstractUdfStreamOperator.java:102) > at org.apache.flink.streaming.runtime.tasks.StreamTask. > openAllOperators(StreamTask.java:372) > at org.apache.flink.streaming.runtime.tasks.StreamTask. > invoke(StreamTask.java:261) > at org.apache.flink.runtime.taskmanager.Task.run(Task.java:609) > at java.lang.Thread.run(Thread.java:745) > > > Is there any way to solve this problem? > > > (The rolling sink function is working based on the processing time and we > need event time for hdfs. Therefore we have choosen bucketing sink > function.) > > > Thank you for your helps. > > > Ozan. >
