I’ve been working on an animation for a couple weeks now, and suddenly I can’t open the level its based in anymore… which is infuriating because its going great otherwise and I’m very near the finish line. I feel like this level just has an incurable disease. Consistently getting:
The level itself involves nested iterations of the same blueprint actor, which houses tons of geometry I’m looping through. The blueprint actor still opens up just fine in the blueprint editor and can drop in new levels easily, and with MORE instances than the cursed level, so it doesn’t appear to be a VRAM issue or a level size/complexity concern. (on an RTX 3090, my hardware tools were showing only 16gb vram occupied when the crashes began, and I’ve pushed it to 22GB vram on more complex animations).
I’ve tried googling this, I’ve tried reinstalling the engine, trashing all temp files, restoring from autosaves of the level before this started happening, creating an entirely new project and re-importing the level and blueprints connected to this… at my wit’s end here.
Has anyone encountered an issue with this directory/file? I’m not a dev so much as an animator, so please speak layman lol.
Current troubleshooting step is attempting to delete anything that could be ‘experimental’ or ‘chaotic.’ IE some niagra systems i’ve got in the cursed level.
Desperate for some help on this. Thank you! Below is the full summary on crash:
Hey there @Filmattic! Welcome back to the community! So this error is seemingly stemming from a chaos physics issue. Are you using the chaos cloth sim or destruction? The assertion is basically involving the mass on the physics object being changed outside of the range it expected. Also you don’t have editor symbols installed so the other portions of the bug report aren’t visible, though the line that is visible gives us a bit of a hint.
The only thing in the project that description might fit were some niagra systems. I deleted those from the project’s file structure and reloaded the project. Same crash when opening the level.
As of this morning, I have gotten the project to a stable place by deleting elements from the nested blueprint actor in reverse chronological order. Once the level opened up without issue, I added back in everything except the most recently deleted object. Still fine. I add back in that one object, and the same crash pops up. So it seems to be something to do with a specific Kitbash3D model. This makes no sense to me whatsoever given the error I’d been getting, as the model’s blueprint has no moving parts or sims in it.
Perplexing, but stable for now. Thanks for your response!
Had this error occur with 5.2. Turned out it was caused by a static mesh’s collision. Removing collision from the static mesh fixed it, now able to compile without errors. (removed collision by opening the static mesh in the editor → clicking “Collision” from the top menu → clicking “Remove Collision”).
Also, initially searched for “error: combine worldspace: zero total mass detected”, which shows up just below the “newmp.mass > (1.e-8f)” error. No results on google, so mentioning it here so the magical algorithms pick it up for the next person.
Thank you for your answer, but I cannot understand exactly what I should look for in the warnings when I receive this error. There were blueprints that I used chaos, I deleted them and it still gives the same error. What causes this error? If I can understand what caused the error, it will be easier to solve it. Thank you for your interest.
I wasn’t too sure what’s it all about. I thought maybe it has to do something with the Chaos plugin? So I’ve tried to turn it off, but it didn’t solve the Error. It was confusing because I wasn’t even using Chaos. And so, I’ve tried to search for Chaos in the details panel and I’ve found this:
This made me realize it probably has to do something with collision!
Simply going to the Actor that was stated “will result in a bad body instance” and changing its ‘Collision Presets’ from ‘BlockAllDynamic’ to ‘NoCollision’ fixed the issue and errors for me!
I think it make sense as if you have collision and also try to scale the object to 0, maybe it will collide with itself, in a bad way, I guess, I think (lol)
Well no,
Bad Body Instance implies that one of the body instances defined on the model itself is bad.
But the warning doesn’t say that, it says that scaling to 0 means you cannot test for collision which makes sense regardless of what physics engine is in use.
Its a warning, not an error.
It seems to me that the Cooker for IOS is overly aggressive on this, converting what is Rightfully a warning into a wrench, and throwing it into the engine’s gears…
Could probably use a Bug Report - even if epic won’t do anything about it.
