Ron,
No change in status on this. The patch I made to apply a 15 second socket timeout on the secondaryServerSocket appears to keep us afloat by releasing the blockage before things get out of control.
If there is ever a Remoting 2.2.3 patch for this, we'd love to have it, but we've actually start planning our next steps assuming there won't be one. In the short term we will look towards trying to maintain longer term connections to stay away from the contention on the callback collection. In the long term we are looking into going to RMI delivery. We need to get to load balanced clustering at some point and your documentation says that excludes the JMS services.
Any ideas you have are appreciated.
Thanks,
David