[
https://issues.apache.org/jira/browse/HDFS-10222?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
James Clampffer updated HDFS-10222:
-----------------------------------
Attachment: HDFS-10222.HDFS-8707.000.patch
Added patch.
The fix has two parts:
1) Always call shutdown() before closing a socket
2) Catch the exceptions that shutdown() may throw. asio converts error codes
from posix functions into std::exceptions, the error codes happening here
aren't really worth caring about. These show up when "delete this" is called
in the continuation after a failed connect. The socket isn't open or closed as
far as asio is concerned so weird things happen. We should plan resource
management better so we can remove "delete this".
> libhdfs++: Shutdown sockets to avoid "Connection reset by peer"
> ---------------------------------------------------------------
>
> Key: HDFS-10222
> URL: https://issues.apache.org/jira/browse/HDFS-10222
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: hdfs-client
> Reporter: James Clampffer
> Assignee: James Clampffer
> Attachments: HDFS-10222.HDFS-8707.000.patch
>
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
