IRRd-Discuss

[Peter Caffin]

Hi Jerry,

On Mon, 6 Nov 2000 gerald@merit.edu wrote:
> > Firstly, I'm looking to using IRRd more as a record management tool for
> > managing DNS records in a similar way to IRRd-based auto-dbm sites. Does
> > IRRd allow the specification and use of other Objects for similar
> > maintennance?
> 
> First let me say that I will probably need more info to answer your
> question as I would need more specifics regarding your goals.
> 
> But assuming you're asking me if you can define new objects (ie,
> DB objects irrd does not currently understand/recognize).  The
> answer is no, not automatically.  You would need to upgrade the
> parser and irrd -- not very easy.
> 
> Anyway, if you can give me a few more detials I can give a better
> answer.

One of the things that I've always liked about dealing with APNIC, when
updating in-addr.arpa and inetnum records, and the various maintainer
and nic-hdl objects, is the email interface and the choice it gives for
authentication mechanisms while being completely transparent; ie, where
something's been changed, it's obvious who made the change and what it
was.

What I'd really like to be able to do is to allow clients to do similar
things but with a range of services such as secondary DNS requests, MX
exchanger requests, etc. It would be a nice devolution of responsibility
to our clients who are (for the most part) ISPs with fairly capable
techs.

One of the useful things about IRRd (as seen from a user point of view)
is the way that it does some checking to see that DNS systems, etc, are
alive and are serving the zones as stated in the request. It works well
for the purposes that IRRd has been written to solve.

However, for the purposes I'm looking at, that sort of checking's not
really needed. I'd be looking at IRRd more as 'a db with a rather nice 
set of interfaces and already present and tested querying tools, etc'
that's not peering any records to any NIC sites.

(It's quite possible that I'm investigating the use of a hammer for a
screwdriver job, but, I'd rather put an existing tool to a different use
where possible rather than reinventing the wheel unnecessarily ;-).)

> > Secondly, I haven't been able to get IRRd v1.2.5, 1.9 or 2.0a compiled.
> > I've installed bison and flex with no obvious hitches. The OS is Solaris
> > 2.5.1 (SPARC). Error output below is v2.0a, v1.9, and v1.2.5. 
> 
> I believe your problems will be solved by using our latest cut.  You
> will find it in the same spot as irrd2.0b-Nov6-2000.tar.gz.

Yep, it solved that particular one. However, it died a bit later with
another :-(.

> If you still get the 'statusfile.o' linker error I would like you
> to cd to ~src/programs/IRRd and do a 'make' and cut and past the output
> for me.

Thanks for your assistance. The output is pasted below.


Peter.

----------------------------------------------------------------------
bash-2.02# cd programs/IRRd/
bash-2.02# make
gcc -g -O2 -Wall -Wno-char-subscripts  -DHAVE_CONFIG_H  -D_REENTRANT
-D_POSIX_PTHREAD_SEMANTICS -DFUNCPROTO -I../../include
-I/usr/local/include -D_REENTRANT -D_POSIX_PTHREAD_SEMANTICS -DFUNCPROTO
-I../../include -I/usr/local/include -c  atomic_trans.c
atomic_trans.c: In function `transaction_dels':
atomic_trans.c:120: warning: implicit declaration of function `snprintf'
atomic_trans.c: In function `build_transaction_file':
atomic_trans.c:257: storage size of `fstats' isn't known
atomic_trans.c:273: warning: implicit declaration of function `fstat'
atomic_trans.c:257: warning: unused variable `fstats'
atomic_trans.c:255: warning: `dbsize' might be used uninitialized in this
function
*** Error code 1
make: Fatal error: Command failed for target `atomic_trans.o'
----------------------------------------------------------------------