Hey Simon,
Thanks for the bug report. This is somewhat odd because I'm pretty sure the hang was taken care of in -3, though I would believe that there could be another cause besides the threading code.
Maybe a look at 'ps auxww | sort -m -k 9 | head' or
'ps auxww | sort -m -k 9 | tail' would be helpful to figure out when the earliest and latest problems were.
I run pyzor through amavis and spamassassin, and though I got hit with the hanging processes the first time round, I've been fine after -3.
Unfortunately, I don't find gdb outputs that readable for python, so I've build a logging function into pyzor, mostly for debugging. If you add --log to whatever call is being made to pyzor, a log file should show up in ~/.pyzor/pyzor.log
A good start might be to let it run for a bit, then grep -C3 the log for Error, or at least look the log over by hand to see if you can spot anything.
Oh, and really the best thing you could do, is to find a way to make pyzor hang from the command line, as that is the easiest to reproduce.
Thanks a lot,
Chris
On 05/01/05 09:21:08, Simon Josefsson wrote:
Package: pyzor Version: cvs20030201-3
I have pretty much the same problem as in:
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=301160
although I am running the latest version. Some of the processes are really old as well:
yxa-iv:~# ps auxww|grep pyzor|wc -l 115 yxa-iv:~# ps auxww|grep pyzor|head jas 20421 0.0 0.2 6116 3028 ? S Apr24 0:00 /usr/bin/python /usr/bin/pyzor check jas 20422 0.0 0.2 6116 3028 ? S Apr24 0:00 /usr/bin/python /usr/bin/pyzor check ...
Any ideas?
I'm using sendmail + spamassassin via spamass-milt, and spamasssin invoke pyzor.
My system track sarge, so everything should be up to date.
Thanks!
yxa-iv:~# gdb /usr/bin/python 20421
