hazelcast-jet
hazelcast-jet copied to clipboard
hazelcast
Distributed Stream and Batch Processing
I have Jet 4.3 running. It contains the `jet-cluster-admin` tool. ``` $ ./jet-cluster-admin -o get-cluster-version No port specified, using 5701. No cluster name specified, using 'jet'. No password specified, using...
IMDG provides https://docs.hazelcast.org/docs/4.0.3/javadoc/com/hazelcast/core/HazelcastInstanceAware.html A class such as Runnable that implements `HazelcastInstanceAware` has this method ``` @Override public void setHazelcastInstance(HazelcastInstance arg0) { this.hazelcastInstance = arg0; } ``` However, this is no...
There are multiple reasons why a submitted job is not running: - cluster is doing migration - some member is shutting down - quorum is not met - snapshot is...
we have `topObservedWm` and `coalescedWm` on both edge and processor level: - `ordinal-N/topOpservedWm`: the highest wm on input N - `topOpservedWm`: the highest of the coalesced WMs on any edge...
Currently max. watermark gap is hardcoded as `1000` however this will cause too many WMs to be emitted when using "nanoseconds" as a time unit.
Usability improvement idea: The Logger sink should use a different logger name. This looks hideous: ``` May 03, 2019 3:53:59 PM com.hazelcast.jet.impl.pipeline.FunctionAdapter$AdaptingProcessor.loggerSink#0 INFO: [192.168.86.54]:5702 [jet] [3.0] 1094=62cf0b26-f859-4f32-844d-678b75883f1e ``` when I...
Currently peek() always logs watermarks, this should be optional
Hello Jet Team, our moderately complex workflow almost exclusively uses `localParallelism(1)` set on `Vertex` level, due to inability to set a general default for the local parallelism level, so we...
As a user I want to be able to control the capacity of concurrent queues that Jet will use to implement DAG edges. Why this is important: 1. The default...
The following test case fails for client: ``` public class JobFailureTest extends TestInClusterSupport { @Test public void name() { BatchSource source = SourceBuilder.batch("test", (context) -> new HashSet()) .fillBufferFn((strings, sourceBuffer) ->...
