I got this error on the console while playing my game:
Assertion failed: Multiple dependency jobs should not be in ScheduleDependencies
However I am not using the job system so I am not scheduling any job. I’ve been working on this project with Unity 2018.1.0b13 since the date it was released and this is the first time I see this message.
Hi George,
Could we please get a report + repro for this?
I’m afraid I don’t know how to reproduce this error… I’ve executed this project hundreds of times and only seen this message once, so it’s extremely hard to reproduce and I have no idea of what triggered it. I don’t know if filling a bug report will be of any help since the people in charge of evaluating the case will have very little chance to reproducing it successfully in a reasonable amount of time, but if you tell me to go ahead I will. In my personal opinion we should spread the word and wait util more evidence can be gathered from different people, if the community is aware of this problem maybe someone can provide the missing piece to solve the puzzle, and avoid giving the QA team a hard time unnecessarily.
The only thing I can suggest is enabling a full stack trace for all messages (using the drop-down options of the Console tab) and seeing if you can get any more information about what is triggering this. You may not be using the jobs system but Unity is.
I kinda figured out that. I didn’t know those options were available, thanks!
It seems you are right. It took almost a month and several hundreds of plays in order to get this:
Assertion failed: Multiple dependency jobs should not be in ScheduleDependencies
0x00000001414F00A9 (Unity) StackWalker::GetCurrentCallstack
0x00000001414F66B6 (Unity) StackWalker::ShowCallstack
0x000000014138F54B (Unity) GetStacktrace
0x0000000140888ADC (Unity) DebugStringToFile
0x0000000140883E66 (Unity) JobQueue::ScheduleDependencies
0x0000000140884746 (Unity) JobQueue::ScheduleJobMultipleDependencies
0x000000014087FC61 (Unity) SyncFences
0x0000000140A876E7 (Unity) TransformChangeDispatch::GetAndClearChangedAsBatchedJobs_Internal
0x0000000140790D39 (Unity) RendererUpdateManager::UpdateAll
0x000000014065A09B (Unity) RenderManager::UpdateAllRenderers
0x000000014094B30A (Unity) InitPlayerLoopCallbacks'::2’::PostLateUpdateUpdateAllRenderersRegistrator::Forward
0x000000014094A1DB (Unity) ExecutePlayerLoop
0x000000014094A2A4 (Unity) ExecutePlayerLoop
0x000000014094CE8C (Unity) PlayerLoop
0x00000001413CAD67 (Unity) PlayerLoopController::UpdateScene
0x00000001413C9C09 (Unity) Application::TickTimer
0x00000001415C42ED (Unity) MainMessageLoop
0x00000001415C65DC (Unity) WinMain
0x000000014271FDA7 (Unity) __scrt_common_main_seh
0x00007FFDB1F01FE4 (KERNEL32) BaseThreadInitThunk
0x00007FFDB21EF061 (ntdll) RtlUserThreadStart
Hi
Please can we work together to solve this problem.
That assert means something has gone wrong internally with the data state of the JobQueue.
Please can you confirm your Unity version - is it still “2018.1.0b13”?
Also please can you confirm what player build you were using 32-bit or 64-bit?
Finally what hardware you are using i.e. CPU when you recreated the problem?
Let me know the best way for us to work together on this problem.
Just to add during all of our automated testing (we have tests which stress test “JobQueue::ScheduleJobMultipleDependencies” a lot) we have never seen this message.
Thanks
Jake
No, I’m using 2018.1.0f2
It is quite difficult to reproduce, I need to run it hundreds of times maybe even a couple of thousands in order to see the problem again. Almost one full month passed since the previos time I saw it, and it could pass another month until I face this message again.
That would be nice.
Again, this happens very rarely, it’s extremely hard to reproduce, so maybe a tiny little detail you didn’t take in to account in your automated tests… something so tiny or so hidden in the code that no one has noticed it yet.
Thank you for the replies they are very helpful.
Please can I bother you for some more information:
Please can you confirm what player build you were using 32-bit or 64-bit? (this is very important it might only happen on for instance the 32-bit player build).
Finally what hardware you are using i.e. CPU when you recreated the problem? (this is important because some older CPUs behave differently to the newest CPUs)
Util now I haven’t seen it on a player build, currently I’m on heavy development and most of the time I play it inside the editor, and is there where I’ve seen this message.
I’m using a DELL Inspiron 5555, the CPU reported by the manufacturer is: AMD Quad Core A8 7410 2.2GHz, and the window task manager reports: AMD A10-8700P Raedon R6, 10 Compute Cores 4C + 6G.
Hope this helps.
Thank you that is very helpful it rules out a lot of potential ideas. Also using 2018.1.0f2 rules out one idea I had.
Back to square one a bit in terms of this assert should not be happening.
Is there anyway to get reproduce this within Unity QA? Or if you can reproduce it then it would be good to break/stop on the assert (I can get you a special build with that change if that helps) and get a core dump for me to analyse.
Just to help - it happens very rarely but when it happens it is always the callstack that says
(Unity) SyncFences
(Unity) TransformChangeDispatch::GetAndClearChangedAsBatchedJobs_Internal
(Unity) RendererUpdateManager::UpdateAll
(Unity) RenderManager::UpdateAllRenderers
That might give more clues as to what is going on.
Just to check anything extra/unique about your project i.e. using new or experimental features like SRP?
I don’t believe I can get an idea of what’s causing the problem any time soon, as far as I’ve been able to see it can happen under almost the same conditions of the previous play, I’ve added, deleted, renamed and replaced a lot of code and the error doesn’t show up, then just a small change and it pops up, but if I run the game under the exact same conditions that caused the error it just doesn’t show up again… by the way, I’m running the game on “auto-pilot” mode, so if I play it twice it gets exactly the same inputs and thus shows exactly the same behavior of the previous play, except for this little error message. I play it for sessions of about 2 to 10 minutes, sometimes I have other things to do and leave the machine alone and hour or two playing the game and I haven’t seen this error, then one day after just a few minutes it throws the exception. It really looks like a random event.
I can do that if you provide me with that special build and give me instructions on how to set up VS to catch this event.
I don’t think I’m doing something special or unique with this project, I wanted to experiment with the ECS,but I haven’t done anything with it yet.
these are that packages I have installed:
Ads 2.0.7
Analytics Library 2.0.16
Analytics Standard Events 1.0.13
Cinemachine 2.1.13
Entities 0.0.11
In App Purchasing 2.0.1
Package Manager 1.9.8
PostProcessing 2.0.5-preview
TextMesh Pro 1.2.3
Just remember one thing, inside my FixedUpdate I have a process that loops through a couple of arrays, since these arrays have always the same items I’m creating them on Awake and destroying them OnDestory, my original intention was to use them inside some jobs, so I declared them as NativeArrays, but for different reasons I never used the JobSystem and the arrays remained declared as NativeArrays. This is the only “out of the normal” I can remember.
I just saw this error for the first time today, googled it and came here. I just updated to Unity Personal 64 bit 2018.1.1f1 yesterday (previous version used was 2017.3.0f3) on Windows 10 64 bit. Also was playing inside of editor when it came up. Here’s what my console said about it:
Assertion failed: Multiple dependency jobs should not be in ScheduleDependencies
0x00000001414EE179 (Unity) StackWalker::GetCurrentCallstack
0x00000001414F4786 (Unity) StackWalker::ShowCallstack
0x000000014138DC3B (Unity) GetStacktrace
0x00000001408885DC (Unity) DebugStringToFile
0x0000000140883966 (Unity) JobQueue::ScheduleDependencies
0x0000000140884246 (Unity) JobQueue::ScheduleJobMultipleDependencies
0x000000014087F751 (Unity) SyncFences
0x0000000140A86D57 (Unity) TransformChangeDispatch::GetAndClearChangedAsBatchedJobs_Internal
0x0000000140A870EA (Unity) TransformChangeDispatch::GetAndClearChangedTransformsForMultipleSystems
0x0000000140B368D3 (Unity) PhysicsManager::SyncTransforms
0x0000000140B2ED38 (Unity) PhysicsManager::InitializeClass'::2’::EarlyUpdatePhysicsResetInterpolatedTransformPositionRegistrator::Forward
0x0000000140949ACB (Unity) ExecutePlayerLoop
0x0000000140949B94 (Unity) ExecutePlayerLoop
0x000000014094C77C (Unity) PlayerLoop
0x00000001413C9457 (Unity) PlayerLoopController::UpdateScene
0x00000001413C82F9 (Unity) Application::TickTimer
0x00000001415C213D (Unity) MainMessageLoop
0x00000001415C442C (Unity) WinMain
0x000000014271EDA7 (Unity) __scrt_common_main_seh
0x00007FFB2B3A2774 (KERNEL32) BaseThreadInitThunk
0x00007FFB2BE30D51 (ntdll) RtlUserThreadStart
Looks the same as Jake’s. My console also printed two more warnings right after this error that I have no clue what they mean. Both just said
“No script asset for GameObjectSelectionItem. Check that the definition is in a file of the same name.”
My PC is running Windows 10 64 bit with an AMD FX-350, GTX 770, and 16g ram.
Thank you for the reports. I am on holiday for a week. When I am back we will work on a special build with more logging and abort/crash instead of the assert to let the debugger work and get a core dump.
Thanks. Warning this could be a complex one to find and get fixed.
You’ll figure it out, I believe in you 
In my experience when a bug is extremely hard to find it is also extremely simple to fix. Good luck and good hunting sir!
Quick note… I just saw this message in 2018.1.1f1
me too (2018.1.1f1)