[
https://issues.apache.org/jira/browse/FLINK-7468?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16287612#comment-16287612
]
ASF GitHub Bot commented on FLINK-7468:
---------------------------------------
Github user zhijiangW commented on the issue:
https://github.com/apache/flink/pull/4559
@NicoK , thanks for suggestions.
I understand your point of wrapping the buffer and backlog together in a
new structure returned by `getNextBuffer()` and it really makes sense for
`PipelinedSubpartition`. But for `SpillableSubpartition`, when it begins to
write the buffer to disk, we can not get the total backlog from that. We can
only get the precise backlog by decreasing one for `getNextBuffer()` and
increasing one for `add(Buffer)` .
So I think we can put the `decreaseStatistics` under the lock which can
cover all the subpartitions.
> Implement sender backlog logic for credit-based
> -----------------------------------------------
>
> Key: FLINK-7468
> URL: https://issues.apache.org/jira/browse/FLINK-7468
> Project: Flink
> Issue Type: Sub-task
> Components: Network
> Reporter: zhijiang
> Assignee: zhijiang
> Fix For: 1.5.0
>
>
> This is a part of work for credit-based network flow control.
> Receivers should know how many buffers are available on the sender side (the
> backlog). The receivers use this information to decide how to distribute
> floating buffers.
> The {{ResultSubpartition}} maintains the backlog which only indicates the
> number of buffers in this subpartition, not including the number of events.
> The backlog is increased for adding buffer to this subpartition, and
> decreased for polling buffer from it.
> The backlog is attached in {{BufferResponse}} by sender as an absolute value
> after the buffer being transferred.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
