tech-userlevel archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: proposal: inetd improvements.

On Sat, Jun 05, 2010 at 12:49:23AM +0200, Martin Husemann wrote:
> On Fri, Jun 04, 2010 at 07:43:25PM +0000, Michael van Elst wrote:
> > That's about what we have now, but a bit more graceful because it allows
> > for some retries and recovers automatically. Doesn't need configuration
> > either :)
> How does it recover automatically from wrong command arguments in the service
> specification?
> I'm pretty sure we are not talking about the same variant of failure here.

Whenever the command is corrected it will work again.

You are right that fixing a typo requires to reload inetd manually.

But in my experience a failing exec of the command is rarely caused
by typos. Usually the specific command isn't installed correctly
(e.g. a library is missing after an upgrade) or resides on a
network filesystem that isn't reachable.

N.B. the same logic also recovers from non-exec failures when
a waiting service exits (and is respawned) rapidly. Something
that a limit to the number of children wouldn't see.

                                Michael van Elst
                                "A potential Snark may lurk in every tree."

Home | Main Index | Thread Index | Old Index