On Mon, 2006-12-04 at 15:40 +0100, Mark Martinec wrote:
> Bill,
>
> > Problem 1: after an 'amavisd reload' or 'amavisd stop' command, not
> > all child processes die. The main server thread exits and some children
> > may exit, but often one or more children hang around. Example:
> ...
> > 31249 ? R 877:28 amavisd (ch4-31249-04)
> > Note that 31249 is still hanging around. A manual 'kill -TERM' will not
> > make it go away. I have to use 'kill -9' to zap it.
>
> Net::Server only uses SIGTERM to kill its child processes on shutdown.
> If a child process does not respond to SIGTERM, a manual SIGKILL
> might be necessary.
>
> It would be interesting to find out what the stuck process was doing.
>
> Running at log level 5 for a while would make it possible to examine
> the last log entries by a process, which need to be manually killed.
>
> Attaching truss or strace to such process might also indicate
> what it is doing.
strace shows both hung processes in this state:
child A: futex(0xb7db4370, FUTEX_WAIT, 2, NULL
child B: futex(0xb9c5f28, FUTEX_WAIT, 2, NULL
I'm running with level 5 debugging now and will followup with the
logfile info for the hung process(es).
-Bill
-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
AMaViS-user mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/amavis-user
AMaViS-FAQ:http://www.amavis.org/amavis-faq.php3
AMaViS-HowTos:http://www.amavis.org/howto/
