[mrtg-developers] Re: IPv6 support for MRTG

Rafael Martinez Torres rafael.martinez at novagnet.com
Mon Apr 21 17:20:21 MEST 2003

Hi all :

Sorry for delay, I have just read your messages on patchs for MRTG-ipv6.

I think we should continue discussion on mrtg-developers list.

 Give me a day to evaluate Lorenzo's patch. It was great to see how we
have agreed on some lines on a first look.

Just a day...I think I can contribute to improve the patch with some

My patch stands for as well as for DNS names as for numeric IPv6
addresses, and also is  backwards compatible...

 But does yours accept public@[2003::14]:161...?

Because they are many SNMP IPv6-only severs , you can try net-snmp ...

This is is a very weak point at mrtg script -not the cfgmaker-, since the
Perl underliyng language will evaluate 3:public@[2003::141]:161
impatiently and makes mrtg crash when no dns names are used. ( Operator
expected at "[" )...

(From file mrtg.cfg...

Target[[2003::14]_3]:			3:public@[2003::14]:161


Again, I haven't tested your patch , just  read the mail above... Give me
a day to read te doc ...

Yours seems to check better IPv6 modules's availability...Acknowledges for
your great work... I had no time to program these constraints...

Write you tomorow...Give me a day...

On Thu, 17 Apr 2003, Lorenzo Colitti wrote:

> Tobias Oetiker wrote:
> > Dear Lorenzo and Valerio,
> >
> > your patch looks fine ... It might be worth noteing thogh, that
> > another group has started working on IpV6 for mrtg ...
> >
> > I was contacted last week by rafael.martinez at novagnet.com about
> > this. I guess it might make sense for you two to compare notes.
> Yes, Simon Leinen had already pointed out Rafael's patch to us:
> http://www.ee.ethz.ch/~slist/mrtg-developers/msg00122.html
> The most important difference between our patch and Rafael's patch is
> backwards compatibility. One of our design goals was that this patch
> could be applied to the main MRTG code base and become, in effect, "the
> next version of MRTG": people could just download and use it without
> changing their configuration files, and then just turn on IPv6 support
> if and when they need it without having to patch or recompile MRTG.
> This means our patch does not require recent versions of Perl, does not
> require IPv6 libraries to be installed (of course, they are required for
> IPv6 to work), and will work exactly the same as non-patched versions if
> IPv6 is not turned on. It also means that the final patch will not
> change the interfaces of SNMP_Session.pm, SNMP_util.pm, and BER.pm, as
> requested by Simon (this will be ready in a day or two).
> Our patch also supports devices that have an IPv6 address but do not
> support SNMP over IPv6 (for example, all Cisco routers). With Rafael's
> patch (if I read it correctly), these would have to be specified by
> their numeric IPv6 address. With ours, such targets can be specified by
> hostname by setting the per-target option IPv4Only to yes.
> However, this also means our patch is longer and more complex than
> Rafael's, as it was developed with different goals in mind.
> Rafael: if you want to have a look at our patch, you can find it at:
> http://brunello.uniroma3.6net.garr.it/mrtg/
> Tobias: as regards getting one of the patches applied to the code, what
> do you think would be the best way to proceed? Should we take the
> discussion to the mailing list? Do you think one of these could be applied?
> Thank you and regards,
> Lorenzo and Valerio

Unsubscribe mailto:mrtg-developers-request at list.ee.ethz.ch?subject=unsubscribe
Help        mailto:mrtg-developers-request at list.ee.ethz.ch?subject=help
Archive     http://www.ee.ethz.ch/~slist/mrtg-developers

More information about the mrtg-developers mailing list