[Mimedefang] mimedefang process dies when system runs out of swap

[Scott.Larnach at ed.ac.uk]

Thanks for the speedy response.

On Tue, 1 Oct 2002, David F. Skoll wrote:

> It is OS-dependent.  Some versions of Linux have a very nasty habit of
> killing random processes when memory gets scarce.  I'm not sure what Solaris
> does.  And it might even be code in Sendmail's libmilter which terminates
> when memory is tight, so it's extremely difficult to control what happens.

I guessed this would be a messy area. Naturally my focus is on
configuring sendmail's and the multiplexor's concurrency so that usage
is kept within the machine's capacity. So this is really for abnormal
circumstances. We have had a couple of nasty collateral spam incidents
here in the last 12 months that have interested me in pushing the mail
system's capacity.

> I do not know why it isn't dumping core.  Solaris experts?

Well, I shall be consulting my own Solaris experts tomorrow. I know they
are familiar with machines that run out of swap so don't anybody go
sinking their time into this just now.

Many thanks, Scott Larnach.