On 02/24/10 18:50, Mike Chesney wrote:
Running Bind 9.6.1-P3
We run authorative DNS for 60k+ zones. One one network where we two
dns servers both running the same hardware on Centos 5.4
We see slow dns responses : example
for i in {1..250}; do dig example.com <http://example.com> @localhost
| grep "Query time:"; done;
Centos is a time-sharing system right. I wonder if your time-share is
up and your simply being scheduled off CPU - the network communication
is an opportunity for the scheduler to do that.
Try adding a sleep and see if your results smooth out - more 0 msec and
less msec total.
for i in {1..250}; do sleep 0.5; dig example.com <http://example.com> @localhost | grep
"Query time:"; done;
Sometimes they'll all come back w/ a 0msec response . But every few
runs we see.
; Query time: 501 msec
;; Query time: 111 msec
;; Query time: 0 msec
;; Query time: 0 msec
; Query time: 0 msec
;; Query time: 0 msec
;; Query time: 0 msec
;; Query time: 0 msec
;; Query time: 0 msec
;; Query time: 0 msec
;; Query time: 0 msec
;; Query time: 0 msec
;; Query time: 0 msec
;; Query time: 0 msec
;; Query time: 1461 msec
;; Query time: 0 msec
;; Query time: 0 msec
;; Query time: 0 msec
;; Query time: 441 msec
;; Query time: 0 msec
;; Query time: 0 msec
;; Query time: 0 msec
This is just a snapshot, most other entries are all 0. This doesn't
happen on any of our other dns servers. Load is pretty low on this
machine around .3 4gb ram. Named consumes about 15% of memory and
4% of cpu. Not sure where to look next.
------------------------------------------------------------------------
_______________________________________________
bind-users mailing list
bind-users@lists.isc.org
https://lists.isc.org/mailman/listinfo/bind-users
_______________________________________________
bind-users mailing list
bind-users@lists.isc.org
https://lists.isc.org/mailman/listinfo/bind-users
