[
https://issues.apache.org/jira/browse/FTPSERVER-303?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12711079#action_12711079
]
David Latorre commented on FTPSERVER-303:
-----------------------------------------
Bug confirmed ...
It seems that the problem is in IoDataConnectionFactory.createDataSocket() ,
when using SSL passive mode we had to do some modifications to workaround a
JVM bug and now we are wrapping a regular socket with a sslSocket:
SSLSocket sslSocket = (SSLSocket) ssocketFactory
.createSocket(serverSocket, serverSocket
.getInetAddress().getHostName(),
serverSocket.getPort(), false);
The last parameter is a boolean "autoclose" (close the underlying socket when
this socket is closed) it seems that we want this to true, and changing the
value to true gets rid of the problem.
Niklas, can you foresee any other problems if we changed autoclose to true ?
> Passive data connection stuck in CLOSE_WAIT.
> --------------------------------------------
>
> Key: FTPSERVER-303
> URL: https://issues.apache.org/jira/browse/FTPSERVER-303
> Project: FtpServer
> Issue Type: Bug
> Components: Core
> Affects Versions: 1.0.0, 1.0.1
> Reporter: Alberto Vazquez
>
> We are using a week old snapshot from Ftpserver (it seems a bit newer than
> the 1.0.1 freeze). One of our clients reported that the data connection was
> being kept open for a long time ... They are using a propietary client its
> name I cannot remember and it was an SSL data connection in passive mode.
>
> After checking this using a windows Filezilla client and a Linux server
> running ftpserver (on top of Sun Java Application Server 9) we noticed that
> the FTP Server had several sockets in CLOSE_WAIT state. At the client side,
> the state was FIN_WAIT_2.
> To my mind , this could be related with the Socket leaking bug that was
> previously reported with SSL passive connections or its workaround although
> it might be SJAS fault or-who-knows. Since we don't have a linux machine
> available for testing I don't know what the real problem is.
>
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
