Repeated queries to the root-servers
Bernhard Schmidt
Bernhard.Schmidt at lrz-muenchen.de
Mon Mar 12 17:17:52 UTC 2007
JINMEI Tatuya / 神明達哉 wrote:
>> Did I understand something wrong? Does anyone have an idea what could
>> cause this behaviour?
> I cannot think of a scenario that would cause the described symptom,
> but I'd monitor the cache content while the 40 second test period
> (e.g., every 10 seconds) to see whether and how the .com referrals and
> glues are cached.
Did that (with 1sec resolution). Method:
while true; do /usr/sbin/rndc dumpdb -all; cp
/var/lib/named/log/named_dump.db /tmp/dump.$(date +%s); sleep 1; done
Result: it just vanishes from one second to the other. Content just
before it is gone (1173718895):
com. 172798 NS a.gtld-servers.net.
172798 NS b.gtld-servers.net.
172798 NS c.gtld-servers.net.
172798 NS d.gtld-servers.net.
172798 NS e.gtld-servers.net.
172798 NS f.gtld-servers.net.
172798 NS g.gtld-servers.net.
172798 NS h.gtld-servers.net.
172798 NS i.gtld-servers.net.
172798 NS j.gtld-servers.net.
172798 NS k.gtld-servers.net.
172798 NS l.gtld-servers.net.
172798 NS m.gtld-servers.net.
grep reveals:
dump.1173718894:com. 172799 NS a.gtld-servers.net.
dump.1173718895:com. 172798 NS a.gtld-servers.net.
dump.1173718963:com. 172730 NS a.gtld-servers.net.
so .com glue is only in three of ~60 dumps, but the TTL is increasing,
so dumpdb seems to be broken. dig -t ns com @127.0.0.1 shows that the NS
record is always in the cache (decreasing TTL), but it does not appear
in the dump.
Anything I'm doing wrong regarding rndc syntax?
Regards,
Bernhard
--
Bernhard Schmidt Netzplanung / IPv6
Leibniz-Rechenzentrum der Bayerischen Akademie der Wissenschaften
Boltzmannstr. 1 D-85748 Garching bei Muenchen
Tel: +49 89 35831-7885 Raum I.2.071 schmidt at lrz.de
More information about the bind-users
mailing list