-
Type:
Bug
-
Status: Resolved
-
Priority:
Minor
-
Resolution: Fixed
-
Affects Version/s: 9.10, 10.10, 11.1-SNAPSHOT
-
Fix Version/s: 9.10-HF34, 10.10-HF09, 11.1, 2021.0
-
Component/s: Streams
-
Release Notes Summary:ComputationRunner processLoop times out without errors in logs
-
Tags:
-
Backlog priority:700
-
Sprint:nxplatform 11.1.12
-
Story Points:3
Running 9.10-HF32 with StreamWorkManager, the ComputationRunner intermittently generates the error below. This appears to be a lastReadTime synchronization issue. The value is updated in methods onPartitionsAssigned and processRecord and accessed in method getTimeoutDuration without protection. Suggest method getTimeoutDuration be replaced as follows to fix the problem:
- protected Duration getTimeoutDuration() { long millis = Math.max(0, System.currentTimeMillis() - lastReadTime); return Duration.ofMillis(Math.min(READ_TIMEOUT.toMillis(), millis);
Error generated:
2019-06-06 01:26:24,014 ERROR [defaultPool-02,in:8352,inCheckpoint:8351,out,0,lastRead:1559784384019,lastTimer:0,wm:204444058777681921,loop:724872,checkpoint] [org.nuxeo.lib.stream.computation.log.ComputationRunner] default: Exception in processLoop: Timeout must not be negative
java.lang.IllegalArgumentException: Timeout must not be negative
at org.apache.kafka.clients.consumer.KafkaConsumer.pool(KafkaConsumer.java:1094)
at org.nuxeo.lib.stream.log.kafka.KafkaLogTailer.pool(KafkaLogTailer.java:196)
at org.nuxeo.lib.stream.log.kafka.KafkaLogTailer.read(KafkaLogTailer.java:147)
at org.nuxeo.lib.stream.computation.log.ComputationRunner.processRecord(ComputationRunner.java:262)
at org.nuxeo.lib.stream.computation.log.ComputationRunner.processLoop(ComputationRunner.java:183)
at org.nuxeo.lib.stream.computation.log.ComputationRunner.run(ComputationRunner.java:142)
at java.util.concurrent.Executors$RunnableAdapter.call(Executors.java:511)
...
Master code has the same issue:
https://github.com/nuxeo/nuxeo/blob/master/nuxeo-runtime/nuxeo-stream/src/main/java/org/nuxeo/lib/stream/computation/log/ComputationRunner.java#L339
