Re: [tlug] Runaway iiimfd

["Ian Wells" <ijw@example.com>]

On 21/02/06, Jim Breen <Jim.Breen@example.com> wrote:
> [Ian Wells (Re: [tlug] Runaway iiimfd) writes:]
> >> On 20/02/06, Jim Breen <Jim.Breen@example.com> wrote:
> >> > For now I have turned the daemon off as I need a browser more than an
> >> > IME, but has anyone heard of this problem or able to suggest anything?
> >>
> >> Not a solution, just a protection, but have you considered putting a
> >> tlimit statement in the iiimfd startup script?  That way it will at
> >> least die without doing unpleasant things to your machine.
>
> Would that work? I thought tlimit was inappropriate to a daemon, which
> ideally runs forever.

Well, yes and no.  You don't want a cpu time limit, for instance, but
a virtual memory limit can't hurt - given the choice between dead
daemon and 2GB daemon, I'd prefer dead...

In bash, 'help ulimit' is actually some use.  (ulimit -v 256000 ;
iiimfd) should do nicely with a 256MB limit, which should be more than
enough for that sort of job, I would think.