Re: [MaraDNS list] MaraDNS 2.0.05 released

Remmy,
	Yes I am the maintainer for maradns on Debian. At the moment if you
just keep doing 'apt-get upgrade' you will never get out of the 1.4.x
series on Debian. (On Ubuntu they messed up the version numbers but the
same is true).

	The 2.x series is packaged however. To get it you have to add
"experimental" to souces in '/etc/apt/sources.list'. Then what you get
is that the maradns package only gives you the authoritative server. The
maradns-deadwood will give you the recursive server but you have to do
all the configuration itself.

	To get the 2.x backwards compatible with 1.4.x cannot be done without
asking the user about IP addresses. So the debconf system needs to be
brought in. Even if I go for the really cheap option of usuing ucf I
need to devote soem time to my other DEbian packages and other work.

	I will of course be picking up 2.0.05 shortly.

On 12/02/12 07:14, Remco Rijnders wrote:
> On Sun, Feb 12, 2012 at 02:08:30AM -0500, Sam wrote in 
> <CAJxgfkR0-yStLDFoAC-2yVG6rXgjsTTGS=A-Ktwfp6XEGVGz9w@mail.gmail.com>:
>>> Something I've been wondering, are you aware of any Linux distributions that
>>> offer MaraDNS 2.* instead of 1.4.* ?
> 
>> Nicholas Bamber has worked really hard and has updated the Debian
>> package to use MaraDNS 2:
> 
>> http://packages.debian.org/source/experimental/maradns
> 
>>> Are there any special things one should be aware of when upgrading an
>>> existing installation to the 2.* series?
> 
>> The big thing to keep in mind when updating from 1.x to 2.0.05 is that
>> MaraDNS and Deadwood can not use the same IP.  MaraDNS only serves
>> local (authoritative) records; all recursion (records on other
>> machines) is handled by Deadwood.
> 
>> I have some hints on updating from MaraDNS 1.x to 2.0 here:
> 
>> http://maradns.org/tutorial/update.html#2.0
> 
> Thanks, that's very useful!
> 
> Nicholas, I am interested in hearing how you are handling existing 
> configurations that used both recursive and authorative on the same IP, as 
> well as any options which are no longer supported. Or are you leaving 
> those for the end user to solve?
> 
> Thanks,
> 
> Remmy

-- 
Nicholas Bamber | http://www.periapt.co.uk/
PGP key 3BFFE73C from pgp.mit.edu