[Mimedefang] defunct defang processes
John Nemeth
jnemeth at victoria.tc.ca
Tue Mar 8 03:58:14 EST 2005
On Jul 28, 11:49am, "David F. Skoll" wrote:
} John Nemeth wrote:
}
} > } That doesn't make too much sense. Are you seeing defunct "mimedefang"
} > } threads, or defunct "mimedefang.pl" Perl processes? If the former, then
}
} > The latter, mimedefang.pl processes. Unlike Linux, NetBSD doesn't
} > show threads as if they are seperate processes.
}
} Wow. That's bad. The only way I can see "defunct" mimedefang.pl
} processes lying around is if the multiplexor is blocked somehow, or if
} it misses a SIGCHLD. Either event should yield massive, almost immediate
} problems -- you shouldn't be able to keep going under those circumstances.
I've minimized it by setting MINSLAVES and MAXSLAVES the same.
Even so, if I don't periodically reap the "defunct" slaves, then after
a day or two the system will run out of memory.
} I'm at a loss to diagnose this without a shell account on a NetBSD system.
It's on a production system, so I can't use that one. The only
non-production system I have at the moment is running the latest
development code, since I've been helping with the PAM integration.
This means that I may replace chunks of the OS or even the entire OS at
a moment's notice.
If you point me at the code involved in tracking slaves, I may do
some work on it myself, or I may setup another NetBSD box for you.
}-- End of excerpt from "David F. Skoll"
More information about the MIMEDefang
mailing list