[
https://issues.apache.org/jira/browse/HADOOP-14560?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
John Zhuge updated HADOOP-14560:
--------------------------------
Attachment: HADOOP-14560.002.patch
Patch 002
* Rebase Alex' PR and attach here in order to run Yetus
* Fix a few potential checkstyle issues, e.g., line length exceeds 80 chars
* Rename the property to {{hadoop.http.socket.backlog.size}} to be consistent
with the variable name
* Move the code around a little bit
+1 LGTM
> Make HttpServer2 backlog size configurable
> ------------------------------------------
>
> Key: HADOOP-14560
> URL: https://issues.apache.org/jira/browse/HADOOP-14560
> Project: Hadoop Common
> Issue Type: New Feature
> Components: common
> Reporter: Alexander Krasheninnikov
> Assignee: Alexander Krasheninnikov
> Labels: webhdfs
> Attachments: HADOOP-14560.002.patch
>
>
> While operating WebHDFS at Badoo, we've faced issue, that hardcoded size of
> socket backlog (128) is not enough for our purposes.
> When performing ~600 concurrent requests, clients receive "Connection
> refused" error.
> We are proposing patch to make this backlog size configurable.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
