UDP client handler shuts down, and BIND stops responding

[Mark Andrews]

In message <BN3PR0601MB1445850E80C9FF6E4C5DF9B4D3A10 at BN3PR0601MB1445.namprd06.prod.outlook.com>, Corby Bennett writes:
> 
> I am running BIND 9.11.0x64 on Windows Server 2012 R2, and hosting over 35K=
>  zones.  Unfortunately, running on a non-Windows OS is not an option at thi=
> s time.
> Occasionally BIND will stop responding to queries.  The service keeps runni=
> ng.  Query requests are still logged, but responses are never received by t=
> he DNS client on the other end.  Restarting the service resolves this until=
>  the next time it happens.  This is what I see in my Default.log after the =
> event:
> 
> 31-Oct-2016 21:45:06.830 general: error: socket.c:2525: unexpected error:
> 31-Oct-2016 21:45:07.142 general: error: unable to convert errno to isc_res=
> ult: 234: More data is available.
> 31-Oct-2016 21:45:07.142 general: error: socket.c:2545: unexpected error:
> 31-Oct-2016 21:45:07.142 general: error: SOCKET_RECV: Windows error code: 2=
> 34, returning ISC error 34
> 31-Oct-2016 21:45:07.142 client: error: UDP client handler shutting down du=
> e to fatal receive error: unexpected error
> 
> I don't see any messages logged that would indicate anything out of the ord=
> inary happening in BIND.  Right before the first error message there are en=
> tries for denied dynamic updates, but that is normal for my environment.  F=
> or now, I have implemented a workaround to automatically restart BIND when =
> this problem is detected, but it is a band-aid and I would prefer a real so=
> lution.
> 
> Has anyone ever seen anything like this before that can offer me some guida=
> nce in troubleshooting?  Or, have I uncovered a genuine bug that I should s=
> ubmit to ISC?
> 
> 
> Thank you,
> Corby
> 
I've opened a ticket for you.
-- 
Mark Andrews, ISC
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742                 INTERNET: marka at isc.org