Segmentation Fault (Core Dumped): Why It Crashes
What is Segmentation fault (core dumped)
- Segmentation fault: the process attempted an invalid memory access and received SIGSEGV, so it was terminated.
- Core dumped: the default action for some fatal signals is to terminate the process and write a core dump (a memory/process-state snapshot) that can be analyzed with a debugger.
Immediate actions that save time
Confirm the signal/exit
echo $?
- full command line, inputs, working directory
- version/build info (– version , package version, commit hash)
- whether it’s inside container/WSL/VM/CI
- If it crashes only once, suspect unstable hardware/overclock, corrupted binaries, or nondeterministic concurrency issues.
Get a useful backtrace
If you have a core file on disk
gdb
set pagination off
bt
bt full
info threads
thread apply all bt full
If your system stores cores via systemd
coredumpctl list –reverse | head
coredumpctl gdb
If you can’t get a core dump
Run under the debugger:
gdb -q –args /path/to/executable [args…]
run
thread apply all bt full
Fast root-cause isolation by scenario
You control the source (C/C++/Rust/Fortran)
1. Rebuild for debugging:
- include symbols, reduce optimization, keep frame pointers (recommended).
2. Reproduce under gdb; identify the first stack frame in your code.
3. If the crash is inside malloc/free/memcpy/strlen, assume earlier memory corruption.
4. Use memory tooling (best ROI):
- AddressSanitizer for out-of-bounds/use-after-free/double-free
- UndefinedBehaviorSanitizer for UB-driven crashes
- Valgrind if sanitizers aren’t possible
5. High-frequency bug classes:
- out-of-bounds read/write
- use-after-free / double-free
- null/dangling pointer dereference
- stack overflow (deep recursion / huge stack arrays)
- data race (nondeterministic crash locations)
Prebuilt app / packaged binary
- temporarily unset LD_LIBRARY_PATH, remove LD_PRELOAD, disable third-party plugins/extensions.
- suspect mismatched glibc/libstdc++/driver stack.
Python: segmentation fault core dumped
Most often, a native extension (.so) issue:
- Reproduce in a fresh venv with minimal dependencies.
- Upgrade/reinstall the suspected wheel/extension.
- Bisect imports until the crash disappears; the last enabled native module is usually the culprit.
- If you own the extension: rebuild with symbols and debug it like C/C++.
Node.js
Common cause: native addons compiled against a different Node/V8 ABI:
- reinstall dependencies, rebuild addons, align Node version to addon expectations.
JVM (Java/Kotlin/Scala)
Most JVM segfaults involve JNI/JNA or low-level runtime/flags:
- remove non-default JVM flags, update JVM, isolate JNI libraries; use the JVM crash log if present.
What fixed looks like
- the original reproducer no longer crashes across repeated runs, and
- the fix is covered by a regression test (or sanitizer run is clean on the reproducer if you can rebuild).
Minimal output to paste into a bug report
- exact command + inputs (or a minimal reproducer)
- version/build id
- OS/kernel/arch + container/VM info if relevant
- thread apply all bt full output (from core or live gdb)
- whether it reproduces with plugins/extensions disabled and with a clean environment
About Author
Bhavik Koradiya is the CEO / Co. Founder of Silver WebBuzz Pvt. Ltd. Having 18+ years Experience in LAMP technology. I have expert in Magento, Joomla, WordPress, Opencart, e-commerce and many other open source. Specialties: Magento, WordPress, OpenCart, Joomla, JQuery, Any Open source.
Related Q&A
Segmentation Fault (Core Dumped): Why It
Pip Command Not Found: Complete Fix Guide Get in Touch With UsSubmitting the form below will ensure a prompt response...
Read More
Git Force Pull: The Safe and
Pip Command Not Found: Complete Fix Guide Get in Touch With UsSubmitting the form below will ensure a prompt response...
Read More
How to Delete a Conda Environment
Pip Command Not Found: Complete Fix Guide Get in Touch With UsSubmitting the form below will ensure a prompt response...
Read More