OMEdit crashes when the simulation fails

[Andrea Bartolini]

When the simulation fails, OMEdit crashes with the following message:

Following error has occurred while retrieving detailed gdb backtrace,

Could not attach to process.  If your uid matches the uid of the target
process, check the setting of /proc/sys/kernel/yama/ptrace_scope, or try
again as the root user.  For more details, see /etc/sysctl.d/10-ptrace.conf
ptrace: Operation not permitted.
/tmp/OpenModelica_andrea/OMEdit/omeditbacktracecommands.txt:3: Error in sourced command file:
No thread selected

it seems the problem is the opening of the result file (attached), if you try to open it from File->Open Result File you can reproduce the problem.

OMEdit 1.13.0~dev-37-g1506f4d
Connected to OpenModelica 1.13.0~dev-231-g1d31692
sysop Ubuntu - bit

[Adeel Asghar]

I can open the result file without any issue.
It might be some other reason.

[Andrea Bartolini]

any suggestion for further investigation ?

[Andrea Bartolini]

for example, if I use the cpp runtime the problem disappears...

[Andrea Bartolini]

more....

for example with cpp I get in the window message:

init | error | <p>SimManager: Nonlinear solver 55 stopped at time 0 with error in algloop solver: <br>
Nonlinear solver failed!</p>stdout | error | <p>
Simulation process failed. Exited with code 1.</p>

and in message browser:

[17] 12:38:25 Scripting Error
Failed to open simulation result /mnt/myData/encData/zz_tmpFolder/OpenModelica/PowerGrids.Electrical.Test.OneBusLineLoad1/PowerGrids.Electrical.Test.OneBusLineLoad1_res.mat: Corrupt header (1)

[Adeel Asghar]

Replying to Andrea.Bartolini:

more....

for example with cpp I get in the window message:

init | error | <p>SimManager: Nonlinear solver 55 stopped at time 0 with error in algloop solver: <br>
Nonlinear solver failed!</p>stdout | error | <p>
Simulation process failed. Exited with code 1.</p>

No idea about this error. Probably someone from cpp runtime developers group can help with this.

and in message browser:

[17] 12:38:25 Scripting Error
Failed to open simulation result /mnt/myData/encData/zz_tmpFolder/OpenModelica/PowerGrids.Electrical.Test.OneBusLineLoad1/PowerGrids.Electrical.Test.OneBusLineLoad1_res.mat: Corrupt header (1)

That's fine because the result file is invalid.

Can you share your model?

If you are using v1.13.0-dev-nightly then don't set the milestone as 1.12.0.

[Adeel Asghar]

Martin there is some issue with the read_matlab4.c.
It doesn't crash on Windows.

Limited backtrace at point of segmentation fault
/home/adeas31/OpenModelica/build/bin/../lib/x86_64-linux-gnu/omc/libOpenModelicaRuntimeC.so(+0x16e97)[0x7fac0b50be97]
/lib/x86_64-linux-gnu/libc.so.6(+0x354b0)[0x7fac0bab44b0]
/home/adeas31/OpenModelica/build/bin/../lib/x86_64-linux-gnu/omc/libOpenModelicaRuntimeC.so(+0x31722)[0x7fac0b526722]
/home/adeas31/OpenModelica/build/bin/../lib/x86_64-linux-gnu/omc/libOpenModelicaRuntimeC.so(omc_matlab4_stopTime+0x19)[0x7fac0b526d29]
/home/adeas31/OpenModelica/build/bin/../lib/x86_64-linux-gnu/omc/libOpenModelicaRuntimeC.so(omc_matlab4_val+0x57)[0x7fac0b526d87]
/home/adeas31/OpenModelica/build/bin/../lib/x86_64-linux-gnu/omc/libomcruntime.so(+0x53b9e)[0x7fac0b7a3b9e]
/home/adeas31/OpenModelica/build/bin/../lib/x86_64-linux-gnu/omc/libOpenModelicaCompiler.so(omc_CevalScriptBackend_cevalInteractiveFunctions3+0x4b5f)[0x7fac0c60142f]
/home/adeas31/OpenModelica/build/bin/../lib/x86_64-linux-gnu/omc/libOpenModelicaCompiler.so(omc_CevalScript_cevalInteractiveFunctions2+0xb24f)[0x7fac0ca54a9f]
/home/adeas31/OpenModelica/build/bin/../lib/x86_64-linux-gnu/omc/libOpenModelicaCompiler.so(omc_CevalScript_cevalInteractiveFunctions+0x406)[0x7fac0ca55926]
/home/adeas31/OpenModelica/build/bin/../lib/x86_64-linux-gnu/omc/libOpenModelicaCompiler.so(omc_CevalScript_ceval+0x20c)[0x7fac0ca56a0c]
/home/adeas31/OpenModelica/build/bin/../lib/x86_64-linux-gnu/omc/libOpenModelicaCompiler.so(omc_Interactive_evaluateExpr+0x29d)[0x7fac0ca346bd]
/home/adeas31/OpenModelica/build/bin/../lib/x86_64-linux-gnu/omc/libOpenModelicaCompiler.so(omc_Interactive_evaluateExprToStr+0x183)[0x7fac0ca34953]
/home/adeas31/OpenModelica/build/bin/../lib/x86_64-linux-gnu/omc/libOpenModelicaCompiler.so(omc_Interactive_evaluate2+0x302)[0x7fac0ca36f72]
/home/adeas31/OpenModelica/build/bin/../lib/x86_64-linux-gnu/omc/libOpenModelicaCompiler.so(omc_Interactive_evaluateToStdOut+0x21f)[0x7fac0ca377ef]
/home/adeas31/OpenModelica/build/bin/../lib/x86_64-linux-gnu/omc/libOpenModelicaCompiler.so(omc_Main_translateFile+0x4db)[0x7fac0c8ff8ab]
Segmentation fault (core dumped)

[Martin Sjölund]

@adeas31: I will fix the crash, but I suggest making OMPlot detect the case when 0==reader->nrows (empty data_2 matrix) because we then don't know the start-time, stop-time, etc. We could of course have some sort of recovery method to try to fix the file until the last time it was synched to disk before the crash, but that's secondary.

[Martin Sjölund]

Fixed in ​PR1937

[Francesco Casella]

I have observed the same behaviour also under Windows. I've started a new nightly build to check if this PR solves the problem

[Francesco Casella]

I confirm that the fix also solved the crash issue under Windows.