[pwrdesign]

hi!

Did a render over night, when I got back this morning all nodes has reached a SL of 11.00 except for one that crashed on SL 8.0.

I stopped the render and now everyone is waiting for others except the last one that is pending
So what should I do now?

can I restart mxcl on my crashed node?
Let it render to SL 0.5 and stop it... Will the merge be completed then?

I dont care about that last node really, I just want the merged MXI from the rest 8 nodes.

If this dosnt work, what should I do then?
I've tryed to manually copy out the cooperative.mxi from each node and merge them in Maxwell Render, but I just get an error message then.

[tom]

can I restart mxcl on my crashed node?
Let it render to SL 0.5 and stop it... Will the merge be completed then?

Yes, it should work. You could also try selecting and stopping that server.

[pwrdesign]

Yes that did work!
Puuuhh

Thats great, I miss the mxi of one node, but it saves my day

[johanruben]

If there is another job in the queue that restarted node will continue on the next job. What to do then?

We have experieced som problems with this lately. The machines have 8 GB of ram on XP 64-bit systems, exquipped with Quads, and should not have any problems rendering. Extremely frustrating. We don't want to babysit all renderings during the night...

Edit: Just experienced this once again - but this time the "final mxi" was replaced with only the light energy from the restarted node, ie SL 0.5. 8x8 hours wasted (and deadline approaching). Does this mean that the result of restarting a failed node depends on where the merged data is contained?

Can this be a timeout problem? The merged MXI file is 2.7 GB and takes a while to merge / write.

/J