Server Stops Responding

[cgrande]

So I think my last thread was in 2005 so bare with me...

I started on Max in Dos then I migrated from Max 5 to Max 8 to Max 9 (vray SP2) and now at 2009 (vray SP3) recently and have been getting the same problem on old and new files. With the older files (max9) it's a matter of setting the threads back to 0 and adjusting some settings to make it perform and look as before but what I keep getting is the host servers not responding after working on the file long enough. I'm basically doing DR and then I'll do something that just kills DR in that file. It happens when I restart the rendering after I tweak something. As soon as that happens I try all sorts of madness. Restarting every service everywhere, resetting max, re-openning but too late no more DR in that file unless I reset, start with a blank file, merge in the model from the corrupt file, adjust everything back then hit render using DR and it works again! I also try using a DR scene that is very simple and it works fine. Load the corrupt one and it doesn't. Never happened with SP2. The worse thing would be you had to restart the service and it worked.

The way I know it's busted is when I hit render and it gets to
Scene transferred to buckwild (one of our render nodes)
but no Server status proceeding that. I just get the local computer rendering then a green error that the node stopped responding.

Before I posted this I spent a few days on the forum and on several projects trying to figure this out. Hopefully someone can point me in the right direction. I can send files if necessary
Thanks!!!

PS there are no other max versions on any of the computers.

FILE
Max 2009
Vray SP3
usually around 8-20megs


WORKSTATION
Windows XP 32
Dual Core 4GB
Max 2009
Vray SP3
running vrayspawner
running VRLservice
No firewalls

HOST SERVER
Windows XP 64
Quad Core 8 GB
Max 2009
Vray SP3
running vrayspawner service
No firewalls

[jacksc02]

There is a new feature in SP3a called "restart slaves on render end" which is in the distributed render settings.
Maybe try turning that on and seeing if that helps.

[cgrande]

Tried that with and without on several files at different times. Leaving that on "restart slaves on render end" is more stable but eventually sh** hits the fan and the file ceases to have the ability to trigger the other host renders to do their job. Again if I don't see them starting up the job after the scene has been sent then I know I'll wait forever with the host showing no CPU usage followed by an error in the vray dialogue window.

edit...When it does stop responding I emmediately test the file by opening a working file that DRs and it works but then go back to the other file and it doesnt so I dont think it's a network or server thing.

edit...Could it be Light Cache or even Incremental Save? As soon as I stopped using them servers always respond. I'm also using vray spawner logged in with rights in each server. I have that in my startup folder and no more service. So far I'm not using incremental save but I'm wondering if I'm going to get the same issue due to something I'm doing wrong with LIGHT CACHE? Passes per cpu, etc?