[LA.pm] Problem with FastCGI under Apache 2

[Geoffrey Young]

Benjamin J. Tilly wrote:
> "Geoffrey Young" <geoff at modperlcookbook.org> wrote:
> 
>>> With unsafe signals, if something makes existing database handles
>>> become invalid, we lose our currently executing pages and have the
>>> possibility of unhappy behaviour until the Apache child dies. 

[snip]

>> hopefully, you know that at least with mp1 you don't need to wait
>> around for the child to die for an unspecified amount of time by using 
>> child_terminate(), which offers a graceful child exit. 

[snip]

> It is good to know that, but I'm not entirely sure how I'm supposed to
> use this knowledge.

I guess it was that first part that stood out, as if you were saying "yeah,
we can detect when the dbh is invalid but then we need to wait around for
the child to die for things to fix themselves."  I was just trying to point
out that if you know your child needs to die you can do it yourself quickly
and gracefully.  but I see now that isn't what you were talking about :)

--Geoff