[
https://issues.apache.org/jira/browse/NIFI-13799?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17884745#comment-17884745
]
ASF subversion and git services commented on NIFI-13799:
--------------------------------------------------------
Commit 2e7a39d2005b0387efe49b846efa354a23b4de76 in nifi's branch
refs/heads/main from David Handermann
[ https://gitbox.apache.org/repos/asf?p=nifi.git;h=2e7a39d200 ]
NIFI-13799 Improved Replicated Cluster Response Handling (#9312)
- Return the remote Response Stream in the Replicated Response for unknown or
large content length values
- Buffered smaller responses
- Addressed code warnings in ThreadPoolRequestReplicator
> Replicated Cluster HTTP Requests should avoid Response Buffering
> ----------------------------------------------------------------
>
> Key: NIFI-13799
> URL: https://issues.apache.org/jira/browse/NIFI-13799
> Project: Apache NiFi
> Issue Type: Improvement
> Reporter: David Handermann
> Assignee: David Handermann
> Priority: Major
> Time Spent: 0.5h
> Remaining Estimate: 0h
>
> HTTP request replication between clustered nodes buffers responses before
> returning to the client. This is acceptable for standard operations such as
> status retrieval and configuration changes, but it can result in heap memory
> exhaustion when attempting to download large file content for local
> evaluation. This issue is specific to clustered deployments because
> standalone instances do not require request replication.
> Buffering the response and deserializing into objects is necessary for
> returning merged status information, but it is not necessary for download
> operations. Refactoring the implementation for more selective buffering based
> on content size and operation type would avoid potential memory issues.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)