Bug report: Error: CAM kernel died unexpectedly.

[Anonymous]

When generating the adaptive toolpath for this model, the cam kernel quits unexpectedly every time. It appears to process most of the way through (80+%?) then dies on this particular path generation.

I have several files like this that worked before the april update, but now do not. 

Previously, I had difficulty with these files sometime causing fusion to crash in sketch mode, or switching back and forth from model to sketch.  The surface details were created from an SVG that was imported that to me does not appear to have an unreasonable number of points/lines.  I have had previous trouble with fusion crashing (and many, many CER submitted) when working with these files in CAM space as well, but there I was confident it as at least being reported.  Now I'm not, as only the CAM kernel crashes.

Running Mac, 10.13.4.  OS Crash logs for camkernel are attached.

[tom3WTRX]

I'm getting a similar issue with a toolpath that generated just fine yesterday:

Error: CAM kernel died unexpectedly.

[Anonymous]

I got curious and poked at the crash dumps.  In the camkernel crash dump, the traceback and load tables clearly show that camkernel is using python 2.7 libraries from the system instead of the version being installed with fusion. It's very unusual to install one version of the libraries, but use the OS versions.  The OS version could (and does) change with each OS update, so you can't be sure which one your application would be using, so common practice is to install the exact one you want with your application.

Fusion's Python:  (~/Library/Application Support/Autodesk/webdeploy/production/9f329cda8799d97ef1cc5e0b1d752a40539ce10f/Frameworks/Python.framework/Versions/3.5) 

Anyway, if this helps Autodesk get to the bottom of this sooner, that's great, if not, "I tried" 🙂

[fusion360EP98K]

Interesting..  I just posted the same issue.. and its on a Mac too so sounds like the same problem..

I'll try doing the CAM on my Laptop.. though thats VERY slow...

Wonder how you set the python path...?

[Anonymous]

.4114 just updated....and it's still broken.

[jandyman]

Same issue here. This must be fixed. I would hate to have to start running under Windows

[fusion360EP98K]

It seems to give more diagnostic info though

I sent it in an email

[Anonymous]

Could you share it here as well? I haven't seen any better diagnostic in this release for my problem... just the "Died unexpectedly" message.

[jandyman]

Just FYI. I took the design that was crashing and ran the CAM on my machine control PC, which is a Windows machine. Is is a PITA, but it worked. So the problem seems to be Mac specific. Which is probably not a surprise to anyone.

[fusion360EP98K]

Debug: This operation depends on operation BU1AdBodyBackUpper12_7mm.

Debug: Initialization elapsed time: 0 ms

Debug: Writing rest surface to '/var/folders/23/f67bgp611pjcb9zzv0nzgbp40000gn/T/Neutron/Fusion360CAM/1/rest192.xbsrf'

Debug: Using CAM kernel from: /Users/clifford/Library/Application Support/Autodesk/webdeploy/production/3c98b13539e71d557b06ffc18cf2b4ef7528dceb/Libraries/Applications/CAM360/camkernel

Error: CAM kernel died unexpectedly.

Debug: Autodesk CAM Kernel 1.5.1 (r44596) STABLE 5fb08299cf9399bd96254e0711f6edc5d6deada2 2018/03/02 16:00

CloseCrossovers distance: 1.0163 pull: 1.0163

Internal error 'closecrossovers too extreme' for expression '0' failed at -auto/freesteel/freesteel/fsimplicitarea.cpp:1512 (#1).

Exception in thread CoreRougherProcessThread:

Traceback (most recent call last):

File "/System/Library/Frameworks/Python.framework/Versions/2.7/lib/python2.7/threading.py", line 810, in __bootstrap_inner

self.run()

File "lib/roughsteel/CoreRougher.py", line 643, in run

File "lib/roughsteel/CoreRougher.py", line 1030, in CoreRougherProcess

File "lib/hsmkernel/AdaptiveClearingOp.py", line 478, in getStockBoundariesZ

File "lib/hsmkernel/AdaptiveClearingOp.py", line 368, in makeZSliceRest

File "lib/hsmkernel/AdaptiveClearingOp.py", line 321, in makeZSliceRestSurfaceImpl

File "lib/hsmkernel/hsmkernelPython.py", line 894, in GenWeaveZProfile

RuntimeError: Internal error (see statement in log).

camkernel(788,0x7fffa543b380) malloc: *** error for object 0x7f8670e24458: incorrect checksum for freed object - object was probably modified after being freed.

*** set a breakpoint in malloc_error_break to debug

Aborted

Debug: Total processing time: 22344 ms

Debug: Total elapsed time: 16173 ms

[Anonymous]

How did you get this output?  I'm not seeing it in the diag package, but could be looking in the wrong place.

[fusion360EP98K]

Click on the red exclamation point...

Then click on the turqiose "Bug" button - "Show debug messages"...

Having said that, the one I just did does not show so much detail...

Debug: This operation depends on operation BU1AdBodyBackUpper12_7mm.

Debug: Initialization elapsed time: 0 ms

Debug: Writing rest surface to '/var/folders/23/f67bgp611pjcb9zzv0nzgbp40000gn/T/Neutron/Fusion360CAM/2/rest192.xbsrf'

Debug: Using CAM kernel from: /Users/clifford/Library/Application Support/Autodesk/webdeploy/production/3c98b13539e71d557b06ffc18cf2b4ef7528dceb/Libraries/Applications/CAM360/camkernel

Error: CAM kernel died unexpectedly.

Debug: Autodesk CAM Kernel 1.5.1 (r44596) STABLE 5fb08299cf9399bd96254e0711f6edc5d6deada2 2018/03/02 16:00

Debug: Total processing time: 22244 ms

Debug: Total elapsed time: 15601 ms

..and another setup the second step worked ok with rest machine.!!..

So I created a simple test model with a rectanglualr hole cut from a rectangluar box... One CAM step with a 12.5mm bit  and a second with a 6mm bit with rest machining turned on and it failed..  with some detail...

Debug: This operation depends on operation Adaptive1.

Debug: Initialization elapsed time: 0 ms

Debug: Writing rest surface to '/var/folders/23/f67bgp611pjcb9zzv0nzgbp40000gn/T/Neutron/Fusion360CAM/3/rest3.xbsrf'

Debug: Using CAM kernel from: /Users/clifford/Library/Application Support/Autodesk/webdeploy/production/3c98b13539e71d557b06ffc18cf2b4ef7528dceb/Libraries/Applications/CAM360/camkernel

Error: CAM kernel died unexpectedly.

Debug: Autodesk CAM Kernel 1.5.1 (r44596) STABLE 5fb08299cf9399bd96254e0711f6edc5d6deada2 2018/03/02 16:00

camkernel(1176,0x7000077cd000) malloc: *** error for object 0x7ff64aa81258: incorrect checksum for freed object - object was probably modified after being freed.

*** set a breakpoint in malloc_error_break to debug

Aborted

Debug: Total processing time: 7383 ms

Debug: Total elapsed time: 7329 ms

[jandyman]

Does anyone get the sense that anyone from AutoDesk is actually monitoring this thread? It doesn't do us any good to talk amongst ourselves if no AutoDesk employee is listening. Is the official place to report bugs? Because this bug seems awfully real. And there is no workaround I've found, other than using the Windows version for cranking the CAM.

[Anonymous]

Depends on your support level.  If you have phone support, that would probably be more visible.

[Anonymous]

Here's mine, same file as I references in the beginning of the thread (occurs in a 2d adaptive path):

Debug: Initialization elapsed time: 0 ms

Debug: Using CAM kernel from: /Users/mikep/Library/Application Support/Autodesk/webdeploy/production/3c98b13539e71d557b06ffc18cf2b4ef7528dceb/Libraries/Applications/CAM360/camkernel

Error: CAM kernel died unexpectedly.

Debug: Autodesk CAM Kernel 1.5.1 (r44596) STABLE 5fb08299cf9399bd96254e0711f6edc5d6deada2 2018/03/02 16:00

pthread_mutex_lock[1]: Undefined error: 0

pthread_mutex_unlock[1]: Undefined error: 0

pthread_mutex_lock[1]: Undefined error: 0

pthread_mutex_unlock[1]: Undefined error: 0

pthread_mutex_lock[1]: Undefined error: 0

pthread_mutex_unlock[1]: Undefined error: 0

pthread_mutex_lock[1]: Undefined error: 0

pthread_mutex_unlock[1]: Undefined error: 0

pthread_mutex_lock[1]: Undefined error: 0

pthread_mutex_unlock[1]: Undefined error: 0

pthread_mutex_lock[1]: Undefined error: 0

pthread_mutex_unlock[1]: Undefined error: 0

Debug: Total processing time: 2796 ms

Debug: Total elapsed time: 2553 ms

[fusion360EP98K]

I only have hobby level license so doubt I have any support. Still great that they do that though. But could do with someone with full support making the contact to perhaps raise the awareness. I assume there would be some professional companies using Macs and having similar issues. I’m surprised it’s limited to OSX since I assume it’s basically similar code doing the number crunching. Just with a different target though stil x86 based.?

[Anonymous]

I'm seeing autodesk tech support actively responding to other threads, but not this one.  If someone has more than student/hobbyist support, would be great to submit this through your channels.

[jandyman]

I hope that isn't because of weak Mac OS support. Somehow I suspect that if this was happening under Windows, all hell would be breaking loose. The moral to this story for now is that I can't count on the Mac version of Fusion360. It's a good enough program that I'll fire up Parallels to run it if need be, but it's a real shame. Presumably like everyone else here, I only have a hobbyist license. I should just be happy that Fusion360 is available for hobbyists for free. And I am.

[Anonymous]

Another update (4116), still broken. Even less information this time.  Also noted that the CAM kernel hasn't been updated in any of these recent updates.  This is really irritating, AUTODESK, please respond in some way and at least acknowledge you've seen the issue. 

Debug: Initialization elapsed time: 0 ms

Debug: Using CAM kernel from: /Users/mikep/Library/Application Support/Autodesk/webdeploy/production/0acd222207d1a04aee6bf377adfb024e9b25c19c/Libraries/Applications/CAM360/camkernel

Error: CAM kernel died unexpectedly.

Debug: Autodesk CAM Kernel 1.5.1 (r44596) STABLE 5fb08299cf9399bd96254e0711f6edc5d6deada2 2018/03/02 16:00

Debug: Total processing time: 9674 ms

Debug: Total elapsed time: 9405 ms

[Anonymous]

My issue is fixed in todays update, .4126.  The notes listed various rest machining, and "mac specific" cam kernel issues.