[Pdns-users] greater number of truncated replies with pdns_recursor?
aadams at esnet.com
Fri Mar 24 19:15:54 UTC 2006
I recently upgraded from snapshot 563 to release 2.9.20. This was done
primarily to fix the round robin mx lookup issue that I had previously
reported. In the time between reporting that issue and upgrading I was
forwarding recursive queries to verizon's dns servers (tacky, I know).
Since I upgraded to 2.9.20 I've been using the pdns_recursor. Evidently
I get more truncated UDP replies via this method than I was getting with
Verizon's servers. In just over a week's time I've seen multiple
Microsoft mail servers start exhibiting the behavior described in this
The work around has been to enable the answering of TCP queries, which I
don't specifically have a problem with. I suppose this is more of a
report than a particular request for assistance.
C:\dig>dig earthlink.net mx @ns2
;; Truncated, retrying in TCP mode. <<<< See what I mean?
; <<>> DiG 9.2.3 <<>> earthlink.net mx @ns2
;; global options: printcmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 18467
;; flags: qr rd ra; QUERY: 1, ANSWER: 20, AUTHORITY: 0, ADDITIONAL: 20
;; QUESTION SECTION:
;earthlink.net. IN MX
;; ANSWER SECTION:
earthlink.net. 1744 IN MX 5 mx9.earthlink.net.
earthlink.net. 1744 IN MX 5 mxf.earthlink.net.
earthlink.net. 1744 IN MX 5 mxb.earthlink.net.
Earthlink's MX record is the only one that I can reliably reproduce the
truncated reply situation with.
This 'truncated' reply causes MS SMTP (IIS/Exchange) to retry the
request in TCP mode. Now that I'm allowing TCP connections to my DNS
servers the behavior should cease. I guess my question becomes, why
does the pdns recursor seem to produce more truncated replies than the
(presumably) bind server run by verizon (220.127.116.11) to whom I was sending
recursive requests before I upgraded to pdns 2.9.20?
E Solutions Corporation
More information about the Pdns-users