[
https://issues.apache.org/jira/browse/HDFS-17758?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17936529#comment-17936529
]
ASF GitHub Bot commented on HDFS-17758:
---------------------------------------
KeeProMise commented on PR #7501:
URL: https://github.com/apache/hadoop/pull/7501#issuecomment-2733571520
Hi, @hfutatzhanghb I think the functionality of the router should not be
reflected in the client code. Moreover, I believe we could either set the
ipc.client.async.calls.max to a very large value, or add a configuration for
asynchronous calls that are not limited by ipc.client.async.calls.max.
> Make method checkAsyncCall not throttle router async rpc request
> ----------------------------------------------------------------
>
> Key: HDFS-17758
> URL: https://issues.apache.org/jira/browse/HDFS-17758
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Reporter: farmmamba
> Assignee: farmmamba
> Priority: Major
> Labels: pull-request-available
>
> In my practice, when enable aysnc router rpc, we must adjust
> {quote}ipc.client.async.calls.max{quote} larger. If not, we will get
> AsyncCallLimitExceededException exception.
> The original purpose of method checkAsyncCall is to prevent Out-of-Memory
> from occurring on the client, we should make it not throttle router async rpc
> request.
> BTW, we use other mechanisms to prevent OOM from occurring on the hdfs
> router. please see HDFS-17713
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
