[
https://issues.apache.org/jira/browse/RATIS-1528?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17496467#comment-17496467
]
GuoHao edited comment on RATIS-1528 at 2/23/22, 4:24 AM:
---------------------------------------------------------
Although the timeout is added in the logical side of NettyClient, it is still
executed on the server side and Netty's Channel is not released. The
workerEventLoopGroup processing capacity is limited and will cause blocking.
I considered adding IdleStateHandler to Netty Pipeline to release requests that
go unanswered for a long time.
What do you think of this change? @Tsz-wo Sze
was (Author: nicholas niu):
Although the timeout is added in the logical side of NettyClient, it is still
executed on the server side and Netty's Channel is not released. The
workerEventLoopGroup processing capacity is limited and will cause blocking.
I considered adding IdleStateHandler to Netty Pipeline to release requests that
go unanswered for a long time.
What do you think of this change?
> Hung after a long time running when streaming writes
> ----------------------------------------------------
>
> Key: RATIS-1528
> URL: https://issues.apache.org/jira/browse/RATIS-1528
> Project: Ratis
> Issue Type: Sub-task
> Reporter: GuoHao
> Assignee: GuoHao
> Priority: Major
>
> I was testing stream recently and found hung after a long run. I wonder if it
> is necessary to add IdleStateHandler on Netty Client and Server
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
