[
https://issues.apache.org/jira/browse/FLINK-10941?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16706921#comment-16706921
]
ASF GitHub Bot commented on FLINK-10941:
----------------------------------------
zhijiangW commented on a change in pull request #7186: [FLINK-10941] Keep slots
which contain unconsumed result partitions
URL: https://github.com/apache/flink/pull/7186#discussion_r238203555
##########
File path:
flink-runtime/src/main/java/org/apache/flink/runtime/io/network/netty/PartitionRequestQueue.java
##########
@@ -134,10 +138,17 @@ public void cancel(InputChannelID receiverId) {
ctx.pipeline().fireUserEventTriggered(receiverId);
}
- public void close() {
+ public void close() throws IOException {
if (ctx != null) {
ctx.channel().close();
}
+
+ LOG.info("Close all {} readers pending for close.",
readersToClose.size());
Review comment:
If we rely on the close request for handling release logic, have you
considered the scenarios of tcp channel shared by multiple tasks?
In this case, if one downstream task have consumed all the upstream's data,
but this tcp channel is also reused by other parallelism tasks, then the close
request will not be sent until all the shared tasks are finished. Further in
common sense, the sub partition can be released as along as network server has
transported it completely, no need to wait all the inqueue buffers are
processed completely by downstream side.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
> Slots prematurely released which still contain unconsumed data
> ---------------------------------------------------------------
>
> Key: FLINK-10941
> URL: https://issues.apache.org/jira/browse/FLINK-10941
> Project: Flink
> Issue Type: Bug
> Components: ResourceManager
> Affects Versions: 1.5.5, 1.6.2, 1.7.0
> Reporter: Qi
> Assignee: Qi
> Priority: Critical
> Labels: pull-request-available
> Fix For: 1.8.0
>
>
> Our case is: Flink 1.5 batch mode, 32 parallelism to read data source and 4
> parallelism to write data sink.
>
> The read task worked perfectly with 32 TMs. However when the job was
> executing the write task, since only 4 TMs were needed, other 28 TMs were
> released. This caused RemoteTransportException in the write task:
>
> org.apache.flink.runtime.io.network.netty.exception.RemoteTransportException:
> Connection unexpectedly closed by remote task manager
> ’the_previous_TM_used_by_read_task'. This might indicate that the remote task
> manager was lost.
> at
> org.apache.flink.runtime.io.network.netty.PartitionRequestClientHandler.channelInactive(PartitionRequestClientHandler.java:133)
> at
> org.apache.flink.shaded.netty4.io.netty.channel.AbstractChannelHandlerContext.invokeChannelInactive(AbstractChannelHandlerContext.java:237)
> ...
>
> After skimming YarnFlinkResourceManager related code, it seems to me that
> Flink is releasing TMs when they’re idle, regardless of whether working TMs
> need them.
>
> Put in another way, Flink seems to prematurely release slots which contain
> unconsumed data and, thus, eventually release a TM which then fails a
> consuming task.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
