unexpected round robin behavior #54

Closed
Habbie opened this Issue Apr 26, 2013 · 2 comments

Projects

None yet

1 participant

@Habbie
Member
Habbie commented Apr 26, 2013

When I repeatedly query for type all on a host that has multiple A records ie:"dig mx1.hotmail.com any" I get the expected round robin behavior.

When I run the following "dig mx1.hotmail.com a" repeatedly the results are always in the same order. This is also true when I run "dig hotmail.com mx".

I have configured my packet-cache with a 0 second ttl, in hopes of resolving the behavior, but without success.

My pdns_server and pdns_recursor were built from code downloaded via subversion when changeset 551 was the most current release.

This was brought to my attention by customer's that were having dificulty sending mail to HotMail. When they use a different resolver (not one of mine) their problems sending to HotMail cease.

If there is more information that I can provide to facilitate this request, please let me know. I am also willing to allow recursion to an ip for testing on one of my DNS servers if it will help.

I appreciate any help that you may be able to offer.

@Habbie Habbie was assigned Apr 26, 2013
@Habbie Habbie closed this Apr 26, 2013
@Habbie
Member
Habbie commented Apr 26, 2013

Author: dan.pdns@hacker.dk
We have the exact same problem here.. Customers cant send to hotmail because the first A record returned for hotmail is down..

Aparantly the recursor always returns the A records numerical sorted:
mx3.hotmail.com. 3121 IN A 64.4.50.179
mx3.hotmail.com. 3121 IN A 65.54.244.200
mx3.hotmail.com. 3121 IN A 65.54.244.72
mx3.hotmail.com. 3121 IN A 65.54.245.72

Where as all other nameservers ive dig'ed responds in random/roundrobin order.

Im setting priority to high on this, since its a huge problem for our customers not being able to send to hotmail.
If there is anything i can do to help speed up a solution, just say the word..

  • Dan Faerch
@Habbie
Member
Habbie commented Apr 26, 2013

Author: ahu
Fixed in commit 567, this was a 2.9.19 regression.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment