Display the Python traceback on a crash
Fault handler for SIGSEGV, SIGFPE, SIGABRT, SIGBUS and SIGILL signals: display the Python traceback and restore the previous handler. Allocate an alternate stack for this handler, if sigaltstack() is available, to be able to allocate memory on the stack, even on stack overflow (not available on Windows).
Import the module and call faulthandler.enable() to enable the fault handler.
The fault handler is called on catastrophic cases and so it can only use signal-safe functions (eg. it doesn’t allocate memory on the heap). That’s why the traceback is limited: it only supports ASCII encoding (use the backslashreplace error handler for non-ASCII characters) and limits each string to 100 characters, doesn’t print the source code in the traceback (only the filename, the function name and the line number), is limited to 100 frames and 100 threads.
By default, the Python traceback is written to the standard error stream. Start your graphical applications in a terminal and run your server in foreground to see the traceback, or pass a file to faulthandler.enable().
faulthandler is implemented in C using signal handlers to be able to dump a traceback on a crash or when Python is blocked (eg. deadlock).
faulthandler is part of Python since Python 3.3: http://docs.python.org/dev/library/faulthandler.html
Example of a segmentation fault on Linux:
$ python >>> import faulthandler >>> faulthandler.enable() >>> faulthandler._sigsegv() Fatal Python error: Segmentation fault Traceback (most recent call first): File "<stdin>", line 1 in <module> Segmentation fault
To install faulthandler module, type the following command:
python setup.py install
Then you can test your setup using the following command:
You need a C compiler (eg. gcc) and Python headers to build the faulthandler module. Eg. on Fedora, you have to install python-devel package (sudo yum install python-devel).
faulthandler module API
There are 4 different ways to display the Python traceback:
- enable(): on a crash
- dump_tracebacks_later(): after a timeout (useful if your program hangs)
- register(): by sending a signal (eg. SIGUSR1). It doesn’t work on Windows.
- dump_traceback(): explicitly
Fault handler state (disabled by default):
- enable(file=sys.stderr, all_threads=False): enable the fault handler
- disable(): disable the fault handler
- is_enabled(): get the status of the fault handler
Dump the current traceback:
- dump_traceback(file=sys.stderr, all_threads=False): dump traceback of the current thread, or of all threads if all_threads is True, into file
- dump_tracebacks_later(timeout, repeat=False, file=sys.stderr, exit=False): dump the traceback of all threads in timeout seconds, or each timeout seconds if repeat is True. If the function is called twice, the new call replaces previous parameters. Exit immediatly if exit is True.
- cancel_dump_tracebacks_later(): cancel the previous call to dump_tracebacks_later()
dump_tracebacks_later() is implemented using the SIGALRM signal and the alarm() function: if the signal handler is called during a system call, the system call is interrupted (return EINTR). It it not available on Windows.
enable() and dump_tracebacks_later() keep an internal reference to the output file. Use disable() and cancel_dump_tracebacks_later() to clear this reference.
Dump the traceback on an user signal:
- register(signum, file=sys.stderr, all_threads=False, chain=False): register an handler for the signal ‘signum’: dump the traceback of the current thread, or of all threads if all_threads is True, into file”. Call the previous handler if chain is True. Not available on Windows.
- unregister(signum): unregister the handler of the signal ‘signum’ registered by register(). Not available on Windows.
Functions to test the fault handler:
- _fatal_error(message): Exit Python with a fatal error, call Py_FatalError() with message.
- _read_null(): read from the NULL pointer (raise SIGSEGV or SIGBUS depending on the platform)
- _sigabrt(): raise a SIGABRT signal (Aborted)
- _sigbus(): raise a SIGBUS signal (Bus error)
- _sigfpe(): raise a SIGFPE signal (Floating point exception), do a division by zero
- _sigill(): raise a SIGILL signal (Illegal instruction)
- _sigsegv(): raise a SIGSEGV signal (Segmentation fault), read memory from NULL (address 0)
- _stack_overflow(): raise a stack overflow error. Not available on all platforms.
register(), unregister(), sigbus() and sigill() are not available on all operation systems.
faulthandler.version_info is the module version as a tuple: (major, minor), faulthandler.__version__ is the module version as a string (e.g. “2.0”).
Version 2.1 (2012-02-05)
- Add an optional chain argument to faulthandler.register()
- Fix faulthandler._sigsegv() for Clang 3.0
- Fix compilation on Visual Studio
Version 2.0 (2011-05-10)
- faulthandler is now part of Python 3.3
- enable() handles also the SIGABRT signal
- Add exit option to dump_tracebacks_later(): if True, exit the program on timeout after dumping the traceback
- Change default value of the all_threads argument: dump all threads by default because under some rare conditions, it is not possible to get the current thread
- Save/restore errno in signal handlers
- dump_tracebacks_later() always dump all threads: remove all_threads option
- Add faulthandler.__version__ attribute (module version as a string)
- faulthandler.version is now a tuple
- dump_traceback_later() to dump_tracebacks_later()
- cancel_dump_traceback_later() to cancel_dump_tracebacks_later()
- sigsegv() to _sigsegv()
- sigfpe() to _sigfpe()
- sigbus() to _sigbus()
- sigill() to _sigill()
- register() and unregister() are no more available on Windows. They were useless: only SIGSEGV, SIGABRT and SIGILL can be handled by the application, and these signals can only be handled by enable().
- Add _fatal_error(), _read_null(), _sigabrt() and _stack_overflow() test functions
- register() uses sigaction() SA_RESTART flag to try to not interrupt the current system call
- The fault handler calls the previous signal handler, using sigaction() SA_NODEFER flag to call it immediatly
- enable() raises an OSError if it was not possible to register a signal handler
- Set module size to 0, instead of -1, to be able to unload the module with Python 3
- Fix a reference leak in dump_tracebacks_later()
- Fix register() if it called twice with the same signal
- Implement m_traverse for Python 3 to help the garbage collector
- Move code from faulthandler/*.c to faulthandler.c and traceback.c: the code is simpler and it was easier to integrate faulthandler into Python 3.3 using one file (traceback.c already existed in Python)
- register() uses a static list for all signals instead of reallocating memory each time a new signal is registered, because the list is shared with the signal handler which may be called anytime.
Version 1.5 (2011-03-24)
- Conform to the PEP 8:
- Rename isenabled() to is_enabled()
- Rename dumpbacktrace() to dump_traceback()
- Rename dumpbacktrace_later() to dump_traceback_later()
- Rename cancel_dumpbacktrace_later() to cancel_dump_traceback_later()
- Limit strings to 100 characters
- dump_traceback_later() signal handler doesn’t clear its reference to the file, because Py_CLEAR() is not signal safe: you have to call explicitly cancel_dump_traceback_later()
Version 1.4 (2011-02-14)
- Add register() and unregister() functions
- Add optional all_threads argument to enable()
- Limit the backtrace to 100 threads
- Allocate an alternative stack for the fatal signal handler to be able to display a backtrace on a stack overflow (define HAVE_SIGALTSTACK). Not available on Windows.
Version 1.3 (2011-01-31)
- Don’t compile dumpbacktrace_later() and cancel_dumpbacktrace_later() on Windows because alarm() is missing
Version 1.2 (2011-01-31)
- Add dumpbacktrace_later() and cancel_dumpbacktrace_later() function
- enable() and dumpbacktrace() get an optional file argument
- Replace dumpbacktrace_threads() function by a new dumpbacktrace() argument: dumpbacktrace(all_threads=True)
- enable() gets the file descriptor of sys.stderr instead of using the file descriptor 2
Version 1.1 (2011-01-03)
- Disable the handler by default, because pkgutil may load the module and so enable the handler which is unexpected
- Add dumpbacktrace() and dumpbacktrace_threads() functions
- sigill() is available on Windows thanks to Martin’s patch
- Fix dump_ascii() for signed char type (eg. on FreeBSD)
- Fix tests.py for Python 2.5
Version 1.0 (2010-12-24)
- First public release
- 2011-01-31: Version 1.2 tested with Python 2.5, 2.6, 2.7, 3.1 and 3.2 on Debian Sid
- 2010-12-24: Tested with Python 2.6, 3.1 and 3.2 on Debian Sid
- 2010-12-24: Tested with Python 2.6 and 3.1 on Windows XP
- minidumper is a C extension for writing “minidumps” for post-mortem analysis of crashes in Python or its extensions: https://bitbucket.org/briancurtin/minidumper/
- tipper: write the traceback of the current thread into a file on SIGUSR1 signal: http://pypi.python.org/pypi/tipper/
- crier: write the traceback of the current thread into a file (eg. /tmp/dump-<pid>) if a “request” file is created (eg. /tmp/crier-<pid>). Implemented using a thread. https://gist.github.com/737056
- Python WAD (Wrapped Application Debugger), not update since 2001: http://www.dabeaz.com/papers/Python2001/python.html
Application fault handlers:
- The GNU libc has a fault handler in debug/segfault.c
- XEmacs has a fault handler displaying the Lisp traceback
- RPy has a fault handler
System-wide fault handlers:
- Ubuntu uses Apport: https://wiki.ubuntu.com/Apport
- The Linux kernel logs also segfaults into /var/log/kern.log (and /var/log/syslog). /proc/sys/kernel/core_pattern contols how coredumps are created.
- Windows opens a popup on a fatal error asking if the error should be reported to Microsoft
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
|Filename, size & hash SHA256 hash help||File type||Python version||Upload date|
|faulthandler-2.1.tar.gz (19.2 kB) Copy SHA256 hash SHA256||Source||None||Feb 5, 2012|
|faulthandler-2.1.win32-py2.7.exe (219.0 kB) Copy SHA256 hash SHA256||Windows Installer||2.7||Feb 5, 2012|