[
https://issues.apache.org/jira/browse/FLINK-13013?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16881199#comment-16881199
]
Piotr Nowojski edited comment on FLINK-13013 at 7/9/19 12:48 PM:
-----------------------------------------------------------------
Merged as c27d0d68dc76ef4818f12296b28d78d6a78d4965 to master.
I'm also expecting this to improve performance by a bit (not have to take
{{requestLock}} per each {{InputGate}} access).
was (Author: pnowojski):
Merged as c27d0d68dc76ef4818f12296b28d78d6a78d4965 to master
> Make sure that SingleInputGate can always request partitions
> ------------------------------------------------------------
>
> Key: FLINK-13013
> URL: https://issues.apache.org/jira/browse/FLINK-13013
> Project: Flink
> Issue Type: Sub-task
> Components: Runtime / Network
> Affects Versions: 1.9.0
> Reporter: Piotr Nowojski
> Assignee: Piotr Nowojski
> Priority: Critical
> Labels: pull-request-available
> Fix For: 1.9.0
>
> Time Spent: 20m
> Remaining Estimate: 0h
>
> Currently {{SingleInputGate}} requests partitions only on the first attempt
> to fetch the data. Before requesting partitions, no data availability
> notifications can get through. This doesn't work well with a new non blocking
> {{InputGate}} interface, since on newly created {{SingleInputGates,}}
> {{InputGate#isAvailable()}} might return not available, and it will be only
> able to switch to available, after first call to
> {{SingleInputGate#pollNext()}}. However this might never happen, since caller
> could be waiting indefinitely on {{SingleInputGate#isAvailable()}}.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
