[
https://issues.apache.org/jira/browse/STORM-3646?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Ethan Li updated STORM-3646:
----------------------------
Description:
Flush only happens on executor main thread.
https://github.com/apache/storm/blob/v2.1.0/storm-client/src/jvm/org/apache/storm/executor/ExecutorTransfer.java#L70-L74
If topologies have their own threads within bolts, and when
{code:java}
topology.producer.batch.size
topology.transfer.batch.size
{code}
are larger than 1, some tuples will be batched and possible never be sent to
downstream.
The default value for them is 1. So it is fine when the configs are not
explicitly changed.
was:
Flush only happens on executor main thread. If topologies have their own
threads within bolts, and when
{code:java}
topology.producer.batch.size
topology.transfer.batch.size
{code}
are larger than 1, some tuples will be batched and possible never be sent to
downstream.
The default value for them is 1. So it is fine when the configs are not
explicitly changed.
> Flush only happens on executor main thread
> ------------------------------------------
>
> Key: STORM-3646
> URL: https://issues.apache.org/jira/browse/STORM-3646
> Project: Apache Storm
> Issue Type: Bug
> Reporter: Ethan Li
> Priority: Major
>
> Flush only happens on executor main thread.
> https://github.com/apache/storm/blob/v2.1.0/storm-client/src/jvm/org/apache/storm/executor/ExecutorTransfer.java#L70-L74
> If topologies have their own threads within bolts, and when
> {code:java}
> topology.producer.batch.size
> topology.transfer.batch.size
> {code}
> are larger than 1, some tuples will be batched and possible never be sent to
> downstream.
> The default value for them is 1. So it is fine when the configs are not
> explicitly changed.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
