[
https://issues.apache.org/jira/browse/FLINK-4545?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15897050#comment-15897050
]
ASF GitHub Bot commented on FLINK-4545:
---------------------------------------
Github user NicoK commented on the issue:
https://github.com/apache/flink/pull/3467
Hi @zhijiangW,
actually, the solution I am working on is to replace the network buffers
parameter by something like "max memory in percent" and "min MB to use". For
this to not create buffer bloat in our network stack, I have started to
implement limited `LocalBufferPool` instances which tune their size based on
the actual number of outgoing and ingoing channels. It is actually not much
more complicated than this and I already started on this in my local branch at
https://github.com/NicoK/flink/tree/flink-4545 - expect a new PR within the
week with more details.
> Flink automatically manages TM network buffer
> ---------------------------------------------
>
> Key: FLINK-4545
> URL: https://issues.apache.org/jira/browse/FLINK-4545
> Project: Flink
> Issue Type: Wish
> Components: Network
> Reporter: Zhenzhong Xu
>
> Currently, the number of network buffer per task manager is preconfigured and
> the memory is pre-allocated through taskmanager.network.numberOfBuffers
> config. In a Job DAG with shuffle phase, this number can go up very high
> depends on the TM cluster size. The formula for calculating the buffer count
> is documented here
> (https://ci.apache.org/projects/flink/flink-docs-master/setup/config.html#configuring-the-network-buffers).
>
> #slots-per-TM^2 * #TMs * 4
> In a standalone deployment, we may need to control the task manager cluster
> size dynamically and then leverage the up-coming Flink feature to support
> scaling job parallelism/rescaling at runtime.
> If the buffer count config is static at runtime and cannot be changed without
> restarting task manager process, this may add latency and complexity for
> scaling process. I am wondering if there is already any discussion around
> whether the network buffer should be automatically managed by Flink or at
> least expose some API to allow it to be reconfigured. Let me know if there is
> any existing JIRA that I should follow.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
