File transfers fail

We have a LAN behind a NAT firewall with a wildfire server on the outside. The wildfire server has the proxy service setup on the standard 7777 port.

While sending files between clients, Spark-Spark works without problems (p2p connections as it transfers the file very fast). Psi (on Windows)-Spark also no problem (quite a bit slower, but that seems to be a problem with Psi, still much faster than the Internet connection, so p2p there as well). Spark-Psi however fails with an error in the Spark client (We don’'t even get a file dialog in Psi) as follows:

Oct 23, 2006 10:22:49 AM org.jivesoftware.spark.util.log.Log error

SEVERE: Error occured during file transfer.

(400)

at org.jivesoftware.smackx.filetransfer.FileTransferNegotiator.negotiateOutgoingTr ansfer(FileTransferNegotiator.java:400)

at org.jivesoftware.smackx.filetransfer.OutgoingFileTransfer.negotiateStream(Outgo ingFileTransfer.java:289)

at org.jivesoftware.smackx.filetransfer.OutgoingFileTransfer.access$200(OutgoingFi leTransfer.java:35)

at org.jivesoftware.smackx.filetransfer.OutgoingFileTransfer$2.run(OutgoingFileTra nsfer.java:207)

at java.lang.Thread.run(Thread.java:613)

Not too informative sadly enough. Also any filetransfers to someone outside of our network (also in a firewalled network) all fails (even if the recipient is running Spark).

Spark uses a method of file transfer that does not conform completely to the file transfer XEPs, when doing this we tried to make things as “backward” compatible as possible but, PSI does not like the way we send the request so that is why it is failing.

Thanks,

Alex

Ok, that would explain why the Spark- fails. But why does the Spark-Spark fail when trying to use a proxy? (I didn’'t see anywhere to configure just a file-transfer proxy in the Spark client, so I guess it tries it automatically).

It does take configuration on the Wildfire side as well. You may want to check the Wildfire forum as there are several threads which discuss how to get the proxy in Wildfire configured properly.

Thanks,

Alex