Hello,
On 10 Nov 2015, at 0:15, James Cloos wrote:
"CH" == Christian Hofstaedtler <z...@debian.org> writes:
CH> Upstream suggested to run 'pdnssec check-zone' on any zone that
CH> shows this behaviour, as they think that "just failing it" should
be
CH> covered by their testsuite.
That helped. Now only one zone fails (even though only two zones had
any errors, and previously all zones failed), and it gives a different
error pattern:
pdns[25904]: Signing thread died because of std::exception: Reading
from socket in Signing Pipe loop: Connection reset by peer
but I cannot tell from the log which zone failed. The AXFR which
initiated just before that log succeeded.
And each of them have the correct (new) SOA values on the secondaries.
So it looks like running check-all-zones changed something.
That doesn’t make any sense :( check-all-zones is a readonly
operation.
Thanks for the quick reply. Given that running check-all-zones seems
to
have cured things (notwithstanding that one std::exception), I think
we
can close this (or at least lower its severity) and I'll followup on
the
upstream mailing list.
Given that it works now, yes, let’s move it to the upstream lists so
we get a bigger set of useful eyes on it.
Kind regards,
--
Peter van Dijk
PowerDNS.COM BV - https://www.powerdns.com/
