Send Simulation to Backburner error

[Streetwise]

I keep getting this error when trying to submit to BB. No idea why.

[Svetlin.Nikolov]

Hmm, Max 2019 sounds like it should be okay. I honestly don't remember a time when I had to submit to Backburner and did not have to uninstall and reinstall it again before it can start working, so as a last resort, maybe try that. Also, make sure the Backburner and 3ds Max versions are the same - if you have installed a new Max, or Maya, it could have messed things up...

[Streetwise]

All of my backburner versions are the same across all nodes. I get the same message with the attached scene. For low res Phoenix simulation work, my workstation is fine. It's when I crank up the resolution that it would be nice to use the other nodes on the farm.

[georgi.zhekov]

Hey,

I can't seem to reproduce it in here. Does it help if you uninstall and then reinstall Backburner?

[Streetwise]

I changed the BB Manager to localhost and that has eliminated the error. So now I can see that .max launches on each node that's running BB Server, but the nodes seem to work on frame 0 and then quit, rather than doing all of the frames. I've looked at the documentation, but don't see where you can tell BB to create the Phoenix liquid meshes for each frame.

[Svetlin.Nikolov]

Hey, so if it quits after 1 frame, this might be due to a wrong output path. Where does your Output cache path point?

[Streetwise]

Hey, so if it quits after 1 frame, this might be due to a wrong output path. Where does your Output cache path point?

The output path points to a drive on my workstation that is accessible by all 4 nodes. I set up a Global Group that consists of 4 nodes and then send the Backburner Simulation jobs to the group. If I understand the functionality correctly: let's say I have 1000 frames to simulate. In order to use all 4 nodes, I send a block of 250 frames to each server by running the Backburner Simulation utility 4 separate times with the different blocks of frames. Is that correct?

So after doing that, Render Nodes 04 and 05 complete their tasks (although 04 complains about licensing after completing its task). Render Nodes 03 and 06 quite after doing "frame 0" but give no errors.

I've double-checked paths and reinstalled Phoenix, disabled IV6, the usual stuff. 03 and 06 are still not having it. No idea why "frame 0" comes up when the frame numbers assigned were much higher.

It's fantastic to be able to send my simulations off to at least 2 of the nodes to free up my workstation!

[Svetlin.Nikolov]

Hey,

Simulations are sequential. Each simulation frame depends on the previous one. Unlike rendering, you can't split the simulation sequence into 4 parts. You could run 4 simulations in parallel with different settings and pick the best looking one, or run 4 completely different scenes. If you can break apart your simulation into separate grids which don't have to interact with one another, then you can send these to different machines, but it has to be the entire sequence.

The remote machines should still be able to write their caches successfully though - what is expected is that only the machine starting from frame 0 would give you the result you like, while the other 3 should simulate, but won't be correct. In case the 4 machines start writing to the same caches, then this might be the reason for your problems. If even the machine starting from frame 0 quits immediately, then if you have installed a Phoenix nightly from around last week or so, I've added some more debug logging to the Phoenix log, so it would help pinpoint the issue - can you check the Phoenix log on the Windows drive /PhoenixFD on any one of the remote machines and see if there are any [ERROR] or [WARNING] messages there, or even send the log over?