[op5-users] Nagios/Merlin NEB-related interaction bug?

[Andreas Ericsson]

On 10/30/2009 09:44 PM, Sean Millichamp wrote:
[ a long, thorough and extremely detailed bug-report ]
>
> At a first pass, the solution seems to be to either re-engineer Merlin's
> initialization sequence to not have to do the register/deregister dance
> or requires a Nagios fix to handle the situation where a deregister was
> called while in the callback function being deregistered.  I haven't
> spent a lot of time thinking about a fix yet.  I think I'm done for the
> weekend :)
>
> Hopefully this makes sense and helps someone else trace what I did.
> I'll be happy to answer any questions that I can.
>

It does indeed. Actually, since we ignore all callbacks except the
EVENTLOOPSTART one, we can safely avoid de-registering the callback
at all, although this will cost us some few cycles every now and
then.

A more appropriate fix would be to make Nagios aware of the fact that
the callback can de-register itself and make it handle the situation
gracefully. Stashing temp_callback->next before actually making the
callback would be one such fix. It also mimics the way callbacks are
handled while being de-registered.

Thanks for the thorough analysis.

-- 
Andreas Ericsson                   andreas.ericsson at op5.se
OP5 AB                             www.op5.se
Tel: +46 8-230225                  Fax: +46 8-230231

Considering the successes of the wars on alcohol, poverty, drugs and
terror, I think we should give some serious thought to declaring war
on peace.