In your case, very likely the two channels were running on different subnets
and the name of the server computer on the clients will then resolve
differently. That's why you had to turn off useIpAddress so the clients could
resolve the server name to the correct ip address and channel.
In my case, I am using two different types of channels within the same
appDomain and there seems to be a confusion on the remoting channel selector
on picking the correct path from client to server and server to client.
Thanks for the input.
--C.J.
[quoted text, click to view] "KidO" wrote:
> I had a similar problem using multiple channels to remote the same object
> across mulitple NICs on the same machine. Accessing the "sub-object" on 1
> channel worked fine, but failed on the 2nd channel. Both channels worked
> perfectly when using methods of the "top-level" object that did not involve
> returning a reference to the "sub-object".
>
> I don't know if this will help in your case, but in mine, I had to set the
> useIpAddress property of each channel to false.
>
> --
> - John
>
>
> "CJ Bourn" wrote:
>
> > I have a remoted "server" that maintains a list of objects that are also
> > remotable. The server maintains two channels, TCP for remote communication
> > and IPC for cross-process communication. Remoting seems to get confused on
> > which channel to use when a remote application accesses the server and
> > requests one of its sub-objects. As soon as the transparent proxy on the
> > client side tries to access the sub-object, it fails with an IPC access
> > failure. However, the client interface was requested through the TCP channel
> > over the network, so why does the new proxy forcibly try to use the IPC
> > channel? I've tried setting the priorities of the channels and such to no
> > avail. I'm baffled and can't seem to resolve it.
> >