I am using libunwind 1.4.0 to collect the stack-trace. It almost always works, but there are cases that I either get unspecified (general) error or no unwind info found when calling unw_get_proc_name. I can walk through the stack-trace in core-dumps (using gdb), and the same binary on the same machine runs without any issues the next time I run the program. Anyone experienced a similar issue and knows how to fix this?
Asked
Active
Viewed 300 times
1
samanca
- 153
- 3
- 11
-
I think all the incidents are happening on RHEL 6.2 and after the process catches a signal. I'm not sure if the presence of a signal is causing the issue, since I have not been able to reproduce the issue with the same signal on the same platform, but it sporadically happens when running automated tests. – samanca May 21 '20 at 18:30
-
Can you post stack trace log here or suspicious code snippet you think? – Bojan May 27 '20 at 23:49