[mrtg-developers] Re: IPv6 support for MRTG

[Lorenzo Colitti]

Rafael Martinez Torres wrote:
> 	Congratulations for the great effort. There are many data
> available on tests, running MRTG on a Win32 platforms against
> several devices...The aproach of
> documents (Tesi.doc)  is inherently academic and brilliant...
> 
> 	shouldn't we focus the patch on a Unix-like approach...?

Hi Rafael,

Thank you for your compliments.

Actually Tesi.doc was only an old version of Valerio's master's thesis, 
which is on IPv6 monitoring in general as well as MRTG on IPv6. It was 
only in the same directory by chance.

Not all of the thesis pertains to MRTG, and as far as I know we didn't 
use MRTG on Win32 at all! The IPv6 patch was developed under Linux.

> 1- Doc:
> -------
> 	- Important : ! I miss a simple README.txt on how to install and
> apply the patches as well as for Perl Patch as for mrtg itself.

Yes, this is coming. It should be added to ipv6.pod in the near future.

> 2 - Running:
> ------------
> 
> 	You have tested on a Win32, I suppose...

Actually, all tests were done on Linux, under various distributions. But 
Win32 shouldn't be a problem, as the patch is designed not to influence 
mrtg and cfgmaker's behaviour in any way unless IPv6 is enabled.

As for IPv6 support under Win32, I'm not sure that the IPv6 libraries 
even work under Windows...

> 		"cfgmaker	public at lisa.consulintel.euro6ix.org"
> 		does not solve to a IPv address...
> 			"cannot solve to IPv address"

Did you specify the --enableipv6 option to cfgmaker as specified in the 
docs? If you don't specify this option, the IPv6 code will not be 
enabled and mrtg and cfgmaker will behave exactly the same as an 
unpatched version.

> 	In order to keep backwards compatibilty,
> 	We shouldn't apply any additional flags to make mrtg run
> 	on IPv6. Just detect the class of address passed to "cfgmaker" . May be
> the error above has to do with this...don't know.

We could no doubt do this, which is equivalent to enabling IPv6 by 
default if the libraries are available. We chose not to do it because 
there might be bugs in our code (or in the IPv6 libraries themselves) 
which could affect IPv4 operation.

We did this because one of the goals was that mrtg and cfgmaker should 
behave in exactly the same way as an unpatched version if IPv6 is disabled.

When the patch is merged and gets some testing we can enable IPv6 
support by default, if people like it. This should not be a problem with 
cfgmaker, as it falls back to IPv4 if IPv6 doesn't work (and of course 
IPv6 is only actually turned on if the libraries are available).


Regards,
Lorenzo

--
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