[
https://issues.apache.org/jira/browse/FLINK-23223?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Yun Gao updated FLINK-23223:
----------------------------
Description:
When the flushAways is enabled (namely set buffer timeout to 0), there might be
cases like:
# The subpartition emit an event which blocks the channel
# The subpartition produce more records. However, this records would be
notified since isBlocked = true.
# When the downstream tasks resume the subpartition later, the subpartition
would only mark isBlocked to false. For local input channels although it tries
to add the channel if isAvailable = true, but this check would not pass since
flushRequest = false.
One case for this issue is https://issues.apache.org/jira/browse/FLINK-22085
which uses LocalInputChannel.
was:
When the flushAways is enabled (namely set buffer timeout to 0), there might be
cases like:
# The subpartition emit an event which blocks the channel
# The subpartition produce more records. However, this records would be
notified since isBlocked = true.
# When the downstream tasks resume the subpartition later, the subpartition
would only mark isBlocked to false. For local input channels although it tries
to add the channel if isAvailable = true, but this check would not pass since
flushRequest = true.
One case for this issue is https://issues.apache.org/jira/browse/FLINK-22085
which uses LocalInputChannel.
> When flushAlways is enabled the subpartition may lose notification of data
> availability
> ---------------------------------------------------------------------------------------
>
> Key: FLINK-23223
> URL: https://issues.apache.org/jira/browse/FLINK-23223
> Project: Flink
> Issue Type: Bug
> Components: Runtime / Network
> Affects Versions: 1.14.0
> Reporter: Yun Gao
> Priority: Major
>
> When the flushAways is enabled (namely set buffer timeout to 0), there might
> be cases like:
> # The subpartition emit an event which blocks the channel
> # The subpartition produce more records. However, this records would be
> notified since isBlocked = true.
> # When the downstream tasks resume the subpartition later, the subpartition
> would only mark isBlocked to false. For local input channels although it
> tries to add the channel if isAvailable = true, but this check would not pass
> since flushRequest = false.
> One case for this issue is https://issues.apache.org/jira/browse/FLINK-22085
> which uses LocalInputChannel.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
