Repository navigation
Python crashes on large number of client connects to socketserver #95204
Description
Crash report
Running the code in this tree:
https://github.com/morphex/smps/tree/59fbc78ebddcc3a68c6ea71200d80683b88b8b7d
Can crash the Python interpreter, with no useful feedback on what's wrong, except that the socketserver must have crashed as connections are refused.
Error messages
Two runs of the smps test; which shows that the number of client connections that crashes Python varies; the first run was after I opened the laptop lid, IIRC:
./test.sh
SMPS running at pid 2941
Starting threads:
0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,
Exception <class 'ConnectionResetError'> (104, 'Connection reset by peer')
Exception <class 'ConnectionRefusedError'> (111, 'Connection refused')
Exception <class 'ConnectionRefusedError'> (111, 'Connection refused')
Exception <class 'ConnectionRefusedError'> (111, 'Connection refused')
Exception <class 'ConnectionRefusedError'> (111, 'Connection refused')
Exception <class 'ConnectionResetError'> (104, 'Connection reset by peer')
Exception <class 'ConnectionRefusedError'> (111, 'Connection refused')
Exception <class 'ConnectionRefusedError'> (111, 'Connection refused')
Exception <class 'ConnectionRefusedError'> (111, 'Connection refused')
Exception <class 'ConnectionRefusedError'> (111, 'Connection refused')
Exception <class 'ConnectionRefusedError'> (111, 'Connection refused')
Exception <class 'ConnectionRefusedError'> (111, 'Connection refused')
Exception <class 'ConnectionResetError'> (104, 'Connection reset by peer')
Exception <class 'ConnectionRefusedError'> (111, 'Connection refused')
Test complete
./test.sh
SMPS running at pid 3048
Starting threads:
0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,
Test complete
morphex@morphex-Latitude-E4310:~/smps$
You can run test.sh with ./test.sh 1000 for example to increase the number of clients generated.
Your environment
Python 3.9.13 (main, May 23 2022, 22:02:02)
[GCC 7.5.0] on linux
Linux morphex-Latitude-E4310 5.4.0-120-generic #136-Ubuntu SMP Fri Jun 10 13:40:48 UTC 2022 x86_64 x86_64 x86_64 GNU/Linux
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 20.04.4 LTS
Release: 20.04
Codename: focal
Package: python3.9-minimal
Status: install ok installed
Priority: optional
Section: python
Installed-Size: 5402
Maintainer: Matthias Klose doko@debian.org
Architecture: amd64
Multi-Arch: allowed
Source: python3.9
Version: 3.9.13-1+bionic1
Depends: libpython3.9-minimal (= 3.9.13-1+bionic1), libexpat1 (>= 2.1~beta3), zlib1g (>= 1:1.2.0)
Pre-Depends: libc6 (>= 2.27)
Recommends: python3.9
Suggests: binfmt-support
Conflicts: binfmt-support (<< 1.1.2)
Description-en: Minimal subset of the Python language (version 3.9)
This package contains the interpreter and some essential modules. It can
be used in the boot process for some basic tasks.
See /usr/share/doc/python3.9-minimal/README.Debian for a list of the modules
contained in this package.
Description-md5: cd370bbd0ff51beae7a757794c49c53c
Cnf-Visible-Pkgname: python3.9
Activity
jepler commented on Jul 24, 2022
From what I could see, in your protocol sending the QUIT request on the client causes the whole server to shut down:
if input.lower().strip() == "quit": DEBUG_PRINT("Exiting") self.server.shutdown()and in def _run_client_test(): you have client.quit().
morphex commented on Jul 24, 2022
Ah, of course. Thanks. :)
terryjreedy commented on Jul 25, 2022
For future reference, the label 'crash' means a hard exit, as with stack overflow. You gave no indication that you say such. Python exceptions are at most a bug. I changed the label for the benefit of anyone searching for true crashes.
Similarly for changing the close reason, which is confusing if one does not know about the drop menu.