Am 05/10/2018 um 16:08 schrieb KatolaZ:
> On Fri, Oct 05, 2018 at 04:19:18PM +0300, Boian Bonev wrote:
>> Hi Eviham,
>>
>> Can you please check now?
>>
>> Recently we suffered a HDD failure and during restore have made a
>> configuration mistake.
>>
>> With best regards,
>> b.
>
> Dear Boian,
>
> thanks a lot for the prompt reply and action. It looks like it is
> working now (meaning that the webserver is serving packages via IPv6
> now). Let's wait for a confirmation from Evilham, and if everything is
> OK I will re-add the mirror to the RR.
>

Indeed fixed, thanks a lot for the very quick response and fix!

We can safely re-add this to the pool.

I am aware this was kinda a weird corner case; that is nice, because
this way we can see that we this whole class of potential issues before
that ever happens :).
I will see that we get this in the checks over next week.

The issues that KatolaZ mentioned will obviously not be caught by that
additional check, but that is out of our control without access to the
mirrors' networks (which we don't have or want).
In any case, I'd trust anyone running such a network to have their own
exhaustive checks in-place, while multiple records for the same name are
common and are something we can pro-actively check.
--
Evilham