Stale cache feature problems

Borja Marcos borjam at sarenet.es
Thu Nov 11 09:54:50 UTC 2021



> On 11 Nov 2021, at 10:40, Blažej Krajňák <blazej.krajnak at gmail.com> wrote:
> 
> Hi,
> 
> št 11. 11. 2021 o 10:28 Borja Marcos <borjam at sarenet.es> napísal(a):
>> First problem: I experienced random SERVFAILS with no apparent reason while i had the feature turned on. I think it
>> especially affected CDNs with multiple chained CNAMES and complex DNS server infrastructures.
>> 
> 
> glad to hear the confirmation of this problem. I experienced the same
> behaviour in our ISP network. Issue  #2982
> https://gitlab.isc.org/isc-projects/bind9/-/issues/2982
> 
> Please, add your conditions and findings to ticket.

I will.

I’ll add what I have for now, but it’s just anecdotal evidence.

I can try to get more debugging information including packet captures. I imagine it has something
to do with an interaction between CDNs (short TTL A records for DNS servers) and making the
wrong decision to use stale records.

It won’t be much of a hassle at home. As it is affecting mostly the Ubiquiti access points I can arrange
for them to use the misbehaving bind, and the rest of the network to use different servers.





Borja.




More information about the bind-users mailing list