This is a discussion on RE: packet too big within the Bind Users forums, part of the DNS and Related Forums category; > -----Original Message----- > From: bind-users-bounce@isc.org [mailto:bind-users-bounce@isc.org] On > Behalf Of ...
|
|||||||
| FAQ | Members List | Calendar | Search | Today's Posts | Mark Forums Read |
|
|||
|
> -----Original Message----- > From: bind-users-bounce@isc.org [mailto:bind-users-bounce@isc.org] On > Behalf Of Jim Reid > Sent: Friday, July 09, 2004 11:37 AM > To: Michael Varre > Cc: bind-users@isc.org > Subject: Re: packet too big > > >>>>> "Michael" == Michael Varre <bind9@kishmish.com> writes: > > Michael> Yes, they are being blocked because they are larger than > Michael> 512 bytes - I just don't understand why they are that > Michael> large. Seems there should be a better explanation than > Michael> just allowing larger packets through via a fixup. > > There is nothing in the DNS protocol that limits answers to 512 bytes. > The string in a TXT record for instance can be up to 64 Kbytes. So it > can't be assumed any answer from the DNS will be less than 512 bytes. > That said, most DNS replies are < 512 bytes to avoid truncated > reponses and retried queries over TCP. However this cannot be assumed > or guaranteed. You have no way of controlling what data other people > put in their zones and therefore how much data their name servers have > to send in a query response. There's even a DNS protocol extension, > EDNS0, which allows for bigger UDP payloads. This will be a Big Win > for things like DNSSEC, ENUM & IPv6 which can make DNS responses much > bigger than they have been in the past. > > If you have a firewall that's blocking DNS payloads of more than 512 > bytes (ie EDNS0 packets or DNS traffic over TCP), it's broken. It's > that simple. Ok, so plain and simple my pix should not be blocking dns packets larger than 512bytes - it is an error on the pix's end. I didn't want to do that unless it were the _correct_ fix - thanks for your help everyone - hopefully the day will get better now :) mv |