So it can't be done is what you're saying.  I haven't really seen any good
arguments as to why it shouldn't be done, but obviously the DJ cronies
aren't going to argue his logic.  It's frustrating for someone like me who
can recognize the many advantages of qmail, even with this little set back
it kills sendmail, but when you run into a feature that seems to be useful
(I'm sure I'm ot theonly one), then you're screwed because Dan says so.

-jeremy

> Jeremy Hansen <[EMAIL PROTECTED]> wrote:
> 
> >My remote concurrency is 500, so that's not a problem.
> 
> 500 is not a magic number guaranteed to be sufficient in all
> applications.
> 
> >In what I've seen
> >and you have to understand, I'm just an admin here, I really have nothing
> >to do with the "quality" of the mail addresses that come through
> >here.  That's a story within itself, but out of 2 million emails, 40% of
> >those on average are deferred.  That's a lot of emails sitting getting
> >retried,
> 
> They're only retried when their schedule says it's time to retry them.
> 
> >I've often seen my remote concurrency consist completely of
> >deferral retries,
> 
> That right there says that your concurrencyremote probably isn't high
> enough.
> 
> >especially of for some reason qmail need to be restarted
> >or something...
> 
> Restarts are exceptional.
> 
> >so if it has to sort through 800,000 mails that may get
> >deferred again, that's wasted time and resources.
> 
> Yeah, avoid restarts.
> 
> >I know I have options, but a deferral host just seems like it would be a
> >"nice thing" to have available.
> 
> Ah, but qmail was engineered. Dan doesn't throw in every feature that
> ``just seems like it would be a "nice thing" to have available''. It's 
> conceivable that a fallback host feature could make sense in some
> applications, but I suspect Dan weighed the pros and cons and decided
> it wasn't worth the effort.
> 
> You can: implement it yourself, switch to a mailer that supports it,
> or consider other options with qmail such as calling qmail-remote
> directly and queuing to a fallback host if that fails.
> 
> -Dave
> 


http://www.xxedgexx.com | [EMAIL PROTECTED]
---------------------------------------------

Reply via email to