
Hi There, I wanted to notify the list that anyone operating recursive resolvers would be wise to flush the following zones from the cache: *.skytv.co.nz *.paymarkclick.co.nz Lastnight the name servers for these where switched to ns1, ns2 and ns3.spark.co.nz. The TTL on the NS record was 24hrs at the time of change so these could still be in cache as the old record. Regards Paul

Hi Paul, Seems strange that a planned change would lead to a call to clear caches. Is there some client behaviour you're seeing that suggests that some caches are holding onto old data for too long? Or was there some unexpected complication in your delegation changes? Are the zones that were being served from the old nameservers meaningfully different from the zones being served from the new ones? It might be helpful to illustrate good and bad behaviour (in terms of DNS responses) to avoid unnecessary cache-clearing. An empty cache is a sad cache. Joe On Aug 9, 2015, at 18:11, paul tinson <paul.tinson(a)gmail.com> wrote: Hi There, I wanted to notify the list that anyone operating recursive resolvers would be wise to flush the following zones from the cache: *.skytv.co.nz *.paymarkclick.co.nz Lastnight the name servers for these where switched to ns1, ns2 and ns3.spark.co.nz. The TTL on the NS record was 24hrs at the time of change so these could still be in cache as the old record. Regards Paul _______________________________________________ NZNOG mailing list NZNOG(a)list.waikato.ac.nz http://list.waikato.ac.nz/mailman/listinfo/nznog

Hi Joe, Good call, i am not asking for a wholesale clear cache. Most resolvers will allow for a particular zone to be flushed. There was an issue with a third party set of name servers, and in the interests of getting some impacted customers online we moved them to the spark name servers and updated the delegation. However the TTL was 24 hours, so some caches may still have an entry for the NS that is having issues and not the spark NS. This was not planned and is an abnormal situation. Regards Paul On Mon, 10 Aug 2015 at 10:31 Joe Abley <jabley(a)hopcount.ca> wrote:
Hi Paul,
Seems strange that a planned change would lead to a call to clear caches. Is there some client behaviour you're seeing that suggests that some caches are holding onto old data for too long? Or was there some unexpected complication in your delegation changes?
Are the zones that were being served from the old nameservers meaningfully different from the zones being served from the new ones?
It might be helpful to illustrate good and bad behaviour (in terms of DNS responses) to avoid unnecessary cache-clearing. An empty cache is a sad cache.
Joe
On Aug 9, 2015, at 18:11, paul tinson <paul.tinson(a)gmail.com> wrote:
Hi There,
I wanted to notify the list that anyone operating recursive resolvers would be wise to flush the following zones from the cache:
*.skytv.co.nz *.paymarkclick.co.nz
Lastnight the name servers for these where switched to ns1, ns2 and ns3.spark.co.nz. The TTL on the NS record was 24hrs at the time of change so these could still be in cache as the old record.
Regards
Paul
_______________________________________________ NZNOG mailing list NZNOG(a)list.waikato.ac.nz http://list.waikato.ac.nz/mailman/listinfo/nznog
participants (2)
-
Joe Abley
-
paul tinson