Re: posible bug in daemontools

Mark Johnson <[email protected]> Mon, 22 Mar 2010 00:13:10 -0500
Newsgroups gmane.network.djbdns,gmane.comp.djb.syslog
Message-ID <[email protected]>
On Mon, Mar 22, 2010 at 12:00 AM, Andy Bradford
<[email protected]> wrote:
> Thus said Mark Johnson on Sun, 14 Mar 2010 16:52:45 CDT:
>
>> Anybody see any potential downside to =A0this change? If not, it will be
>> in zinq-daemontools-0.06.
>
> The =A0patch =A0seems =A0to =A0correct =A0the =A0behavior, =A0however, =
=A0as =A0previously
> mentioned, it's =A0debatable whether or not =A0it is desirable to =A0rest=
art a
> critical process =A0when the system is =A0out of memory. I'm =A0not sure =
about
> other resource starvation situations (like too many open files).
>
> Andy

True.  Hmmmm.  What about something like the quadratic backoff from
qmail?  Would that strike a better balance between the two extremes?

I think some sort of failed fork recovery is necessary, otherwise on
the rare occasion when the OOM killer actually whacks the memory hog
causing the problem, you're left with a functional system with a down
service.  8)