Hello mono folks,
after the success of using mono-service with a ServiceProcess-based
application (thanks again J�rg for your remark - yes indeed, there are
problems with service shutdown [but it *ends* somehow...]), I have come
across a weird show-stopper related to remoting/network/??.
Unfortunately, I don't really have a clue on this:
There's basically a client/server situation with the client being a
Windows application calling a server via TcpChannel on Linux/mono. The
remoted singleton object has a method that returns a class that acts
like a CAO. This works very well with my SuSE 9.1 based test
environment. For some reason, it does not work with the server running
on a freshly installed SuSE Linux Enterprise Server 9 (additionally
installed pkg-config and mono-complete RPMs there). My Windows client
gets the proxy instances of the remote objects, but the first call to
the actual worker object fails - with a SocketException, telling me the
destination host is unreachable!? Huh!
Now, equipped with my limited skillset of network analysis, I took the
challenge. Ping works either way using the IP, host name, fully
qualified name. Nslookups (forward and reverse) return the proper
results. I would be glad to learn about (a) what this could be or (b)
how to further tackle down the issue. It seems that a return from the
server side has a wrong URL, perhaps one that points to a localhost
address or so? Might WINS be an issue here (as there's a Samba server on
the PC that works)? What else? There are no firewalls on either side. I
can even telnet the server on the port in question and see there's
"something going on there".
Thanks for your remarks,
Helge
_______________________________________________
Mono-list maillist - [email protected]
http://lists.ximian.com/mailman/listinfo/mono-list
