[PATCH] Protect inetd and cron from being killed under high-pressure swapping

[Gordon Tetlow]

On Thu, Nov 19, 2009 at 10:26 PM, Peter Jeremy <peterjeremy at acm.org> wrote:

> On 2009-Nov-17 21:35:34 +0000, Rui Paulo <rpaulo at gmail.com> wrote:
> >> At Mon, 16 Nov 2009 16:02:08 +0100,
> >> Attilio Rao wrote:
> >>>
> >>> This simple patch protects inetd and cron from being killed under
> >>> high-pressure swapping systems:
> >>>
> http://www.freebsd.org/~attilio/Sandvine/STABLE_8/madvised/madvised.diff<http://www.freebsd.org/%7Eattilio/Sandvine/STABLE_8/madvised/madvised.diff>
>
> >Yes, the idea is good, but should we really trust inetd and crond that
> much ?
> >Are there other daemons that do this?
>
> As an alternative approach, how about placing a wrapper process around
> them which will restart them if they die?  init(8) unofficially provides
> this (I've used it in the past) - maybe we should formalise this.
>

 Restarting cron can be dicey. @reboot tasks run and could cause problems. I
think protecting certain important processes from being killed in low memory
situations is a smart idea.