logging.handlers.SysLogHandler with TCP not working on rsyslog5.8

[ghost]

BPO	27373
Nosy	@vsajip, @matrixise
Files	fail.log: the tcpdump result of sending message with TCP

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = <Date 2016-08-06.16:25:28.514>
created_at = <Date 2016-06-23.09:04:31.770>
labels = ['invalid', 'type-bug', 'library']
title = 'logging.handlers.SysLogHandler with TCP not working on rsyslog5.8'
updated_at = <Date 2016-08-06.16:25:28.513>
user = None

bugs.python.org fields:

activity = <Date 2016-08-06.16:25:28.513>
actor = 'vinay.sajip'
assignee = 'none'
closed = True
closed_date = <Date 2016-08-06.16:25:28.514>
closer = 'vinay.sajip'
components = ['Library (Lib)']
creation = <Date 2016-06-23.09:04:31.770>
creator = '\xe5\x90\xb4\xe5\x86\x89\xe6\xb3\xa2'
dependencies = []
files = ['43518']
hgrepos = []
issue_num = 27373
keywords = []
message_count = 4.0
messages = ['269112', '269118', '272081', '272082']
nosy_count = 3.0
nosy_names = ['vinay.sajip', 'matrixise', '\xe5\x90\xb4\xe5\x86\x89\xe6\xb3\xa2']
pr_nums = []
priority = 'normal'
resolution = 'not a bug'
stage = 'resolved'
status = 'closed'
superseder = None
type = 'behavior'
url = 'https://bugs.python.org/issue27373'
versions = ['Python 2.7']

[ghost]

logging.handlers.SysLogHandler with TCP not working on rsyslog5.8, but newer version rsyslog8.19 is fine.

I use SysLogHandler with TCP to send message to remote rsyslog server failed, but success with bash below:

echo "<133>$0[$$]: Test syslog message from Netcat" | nc -w1 -t 192.168.1.23 514

this is my script:
#!/usr/bin/env python
# -- coding: utf-8 --
import logging
import logging.handlers
import socket
import syslog

my_logger = logging.getLogger('MyLogger')  # logger is singleton
my_logger.setLevel(logging.DEBUG)

if __name__ == '__main__':

    handler = logging.handlers.SysLogHandler(address=('192.168.1.23', 514), facility="auth", socktype=socket.SOCK_STREAM)
    my_logger.addHandler(handler)
    my_logger.critical('this is critical')

    handler.flush()
    handler.close()
    my_logger.removeHandler(handler)

CentOS version is 6.6
Python version is 2.7.10

This is easy to repeat on my serval CentOS Server. And I attach tcpdump result of script and bash.

[vsajip]

If it works fine with rsyslog 8.19 but fails with rsyslog 5.8, what exactly is logging doing wrong? How is rsyslog configured in the two cases? How exactly was the fail.log generated (I know you said tcpdump was used, but can you provide more detail)?

I'm not sure I'll be able to reproduce this, so you may need to identify what you think logging is doing wrong (there have been no changes in SysLogHandler for quite a while, AFAIK).

[vsajip]

If no more information is available, I will close this issue soon as "not a bug".

[matrixise]

For me, it's an issue with the management of the protocol in rsyslog and not with the logging library.

You can close the issue.