| Newsgroups |
gmane.comp.java.jxta.user |
| Message-ID |
<[email protected]> |
Hi,
We are running in a problem that might be a bug on the JXTA side. We have a setup in which each JXTA peer runs several threads. Each of these threads simulates a peer running a protocol that we're testing. Each simulated peer communicates with other simulated peers over JXTA sockets.
We are running our program in a local network.
We are now running into the following problem:
Peer A starting successfully, Running simulated Peer A1. Opening JXTA server socket.
Peer B starts one simulated peer (B1). B1 gets initial introduction data from A1. This means: it finds A1's JXTA socket advertisement and connects successfully to the JXTA socket of A1 and retrieves data from there. After communication we close the socket.
Same for new simulated peers B2 through B5. B2 through B5 are started sequentially with a delay of 15 seconds each.
When B6 enters the scene, things change. B6 does manage to discover the A1 pipe advertisement, however, it does not manage to connect to A1.
Then B7, B8 etc. don't even manage to obtain the A1 socket advertisement. End of story.
Things look like there is some leak. We checked that we close all sockets. In any case, however, if we come with a new machine C and running a simulated peer C1 on it, we might get some simulated peers from machine C connected to A1.
Does anyone of you have an idea? If this is the wrong mailing list, please point us to another one. The developer list?? If we should Read some parts of TFM we have overlooked, we would be glad to be pointed to these parts.
Cheers,
Lutz Lindner & (Dr.) Wolfgang Müller