Python logging call crashes at usesTime - python-3.x

Question

Every once in 50-100 calls to the logger, the program crashes with the following trace messages:
Mar 20 07:10:14 service.bash[7693]: Fatal Python error: Cannot recover from stack overflow.
Mar 20 07:10:14 service.bash[7693]: Current thread 0x76fa3010 (most recent call first):
Mar 20 07:10:14 service.bash[7693]: File "/usr/lib/python3.5/logging/__init__.py", line 381 in usesTime
Mar 20 07:10:14 service.bash[7693]: File "/usr/lib/python3.5/logging/__init__.py", line 537 in usesTime
Mar 20 07:10:14 service.bash[7693]: File "/usr/lib/python3.5/logging/__init__.py", line 569 in format
Mar 20 07:10:14 service.bash[7693]: File "/usr/lib/python3.5/logging/__init__.py", line 831 in format
Mar 20 07:10:14 service.bash[7693]: File "/usr/lib/python3.5/logging/__init__.py", line 981 in emit
Mar 20 07:10:14 service.bash[7693]: File "/usr/lib/python3.5/logging/__init__.py", line 856 in handle
Mar 20 07:10:14 service.bash[7693]: File "/usr/lib/python3.5/logging/__init__.py", line 1488 in callHandlers
Mar 20 07:10:14 service.bash[7693]: File "/usr/lib/python3.5/logging/__init__.py", line 1426 in handle
Mar 20 07:10:14 service.bash[7693]: File "/usr/lib/python3.5/logging/__init__.py", line 1416 in _log
Mar 20 07:10:14 service.bash[7693]: File "/usr/lib/python3.5/logging/__init__.py", line 1280 in info
Mar 20 07:10:14 service.bash[7693]: File "*****************_app/base.py", line 63 in log_this
Any idea what could be causing this crash?
Don't see similar or other logging calls elsewhere in the program crashing it.
Here is the stack of the calls made to the logger:
self.info("cs={} miso={} mosi{} clk{}".format( self.csPin, self.misoPin, self.mosiPin, self.clkPin))
|
self.log_this("info", msg)
|
self.log.info(msg)
The logger is setup in the base class initialization routine in the following way:
# Global logger is declared as a class attribute
cls.log = logging.getLogger(cls.args["app"])
c_handler = logging.StreamHandler()
f_handler = logging.handlers.RotatingFileHandler(
cls.args["--log"],
maxBytes=(10**6)*int(cls.args["--logsize"]), # CLI is in MB
backupCount=1)
# Create handlers
if cls.args["--debug"]:
cls.log.setLevel(logging.DEBUG)
c_handler.setLevel(logging.DEBUG)
f_handler.setLevel(logging.DEBUG)
else:
cls.log.setLevel(logging.INFO)
c_handler.setLevel(logging.INFO)
f_handler.setLevel(logging.INFO)
# Create formatters and add it to handlers
c_format = logging.Formatter('%(name)s: %(levelname)s: %(message)s')
f_format = logging.Formatter('%(asctime)s: %(name)s: %(levelname)s: %(message)s')
c_handler.setFormatter(c_format)
f_handler.setFormatter(f_format)
# Add handlers to the logger
cls.log.addHandler(c_handler)
cls.log.addHandler(f_handler)
cls.log.info("Logger initialized")
Thank you.

Answer 1

Are you using Windows and switching to an UTF8 locale within Python? There is a Python bug in this specific scenario (similar issue: https://bugs.python.org/issue36792). Here's a minimal code sample that reproduces the bug on my machine:
import locale
import logging
handler = logging.StreamHandler()
handler.setFormatter(logging.Formatter('%(asctime)s'))
logging.getLogger().setLevel(logging.DEBUG)
logging.getLogger().addHandler(handler)
logging.info(1)
locale.setlocale(locale.LC_ALL, 'en_US.UTF-8')
logging.info(2)
Output:
2019-10-25 21:20:15,657
Process finished with exit code -1073740940 (0xC0000374)
If you are unsure if this might be related to your problem, try adding the following line in front of your call to the logging module and see if it solves the problem:
locale.setlocale(locale.LC_ALL, 'en_US')