Re: cfexecd locks

[Alec H. Peterson]

--On Wednesday, September 24, 2003 7:02 PM +0200 Mark.Burgess@iu.hio.no 
wrote:

> This is sort of correct. You should only have one instantiation of the
> daemon running hence the lock. But it is possible that the process dies
> without releasing its lock, so cfexecd automatically tried to ensure
> consistency. This is not perfect behavour, but it is predictable.
>
> I cannot think of a better solution that will work in the absence
> of human intervention.

How about if in addition we were to have a 'daemon' lock acquired after the 
fork (and thus with the appropriate process ID)?  That way cfexecd could be 
restarted simply by running it again.

Alec

p7sMT4JMOnt2R.p7s
Description: S/MIME cryptographic signature