steveniemitz opened a new issue, #22080:
URL: https://github.com/apache/beam/issues/22080

   ### What happened?
   
   https://issues.apache.org/jira/browse/BEAM-14187 attempted to fix an NPE in 
IsmReaderImpl, however it doesn't seem like this is completely fixed.
   
   We're seeing errors like:
   ```
   Caused by: java.lang.NullPointerException
        at 
org.apache.beam.vendor.guava.v26_0_jre.com.google.common.base.Preconditions.checkNotNull(Preconditions.java:877)
        at 
org.apache.beam.vendor.guava.v26_0_jre.com.google.common.base.Optional.of(Optional.java:103)
        at 
org.apache.beam.runners.dataflow.worker.IsmReaderImpl.initializeForKeyedRead(IsmReaderImpl.java:566)
        at 
org.apache.beam.runners.dataflow.worker.IsmReaderImpl.overKeyComponents(IsmReaderImpl.java:200)
        at 
org.apache.beam.runners.dataflow.worker.SideInputTrackingIsmReader.overKeyComponents(SideInputTrackingIsmReader.java:104)
        at 
org.apache.beam.runners.dataflow.worker.IsmSideInputReader.findAndStartReaders(IsmSideInputReader.java:707)
        at 
org.apache.beam.runners.dataflow.worker.IsmSideInputReader.access$500(IsmSideInputReader.java:107)
   ...
   ```
   
   
https://github.com/apache/beam/blob/a3b3182e38fe6b2152f371d4232ddc5d22feed71/runners/google-cloud-dataflow-java/worker/src/main/java/org/apache/beam/runners/dataflow/worker/IsmReaderImpl.java#L566
   
   I can't completely follow the logic going on in `initializeForKeyedRead`, 
but it seems like if multiple threads race on computeIfAbsent, the losers will 
all throw an NPE?
   
   ### Issue Priority
   
   Priority: 1
   
   ### Issue Component
   
   Component: runner-dataflow


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to