On Tue, Dec 15, 2020 at 8:11 PM Jay R. Ashworth via mailop <
[email protected]> wrote:

> ----- Original Message -----
> > From: "John Levine via mailop" <[email protected]>
>
> > In article <[email protected]> you write:
> >>things break, it happens...
> >>
> >>but why 5xx (vs 4xx) in this case?
> >
> > Because it's broken.
> >
> > HTH, HAND,
> > John
>
> Sure, I haven't had to sweep up behind John for a while, why not?
>
> "Because the send-mail cluster and the receive-mail cluster aren't both
> properly connected to the same auth server, so the receive-mail cluster
> can't
> tell those are valid mailboxes... even though presumably it *has the
> physical
> mailboxes handy*.
>

Sometimes, I wish they'd publish at least one actual post-mortem from one
of these outages,
they are usually amazing pieces of rube-goldbergian choose your own
adventure.  Converting those
into public incident reports that are understandable to a much larger
audience definitely loses a lot.
Of course, you'd also need to publish a glossary to even explain what all
the bits were that were involved.

Complex systems break in amazing ways, and the code archaeology of how you
even ended up with
that system is often lost to the sands of time and "I thought it was a good
idea then".

Brandon
_______________________________________________
mailop mailing list
[email protected]
https://list.mailop.org/listinfo/mailop

Reply via email to