On fre, 2007-09-28 at 02:00 -0500, Chris Abbey wrote:
> Hmm. So here's my problem with this. On the client side, I might not know 
> that I 
> need to use IPv6 to get to the server. Say all I know is that I need to send 
> traps to entrapment.example.com. Hand that name off to DNS and it only comes 
> back with an AAAA record (an IPv6 address). How is a script or a user calling 
> snmptrap to report something or query something supposed to know that they 
> need 
> to specify "udp6:" for this address?

Yes, this is a problem. For the special case of IPv4/IPv6 I think the
best course forward would be to merge the two drivers into one, but in
the general case I do not know how to do this, but on the other hand I
think merge of IPv4 and IPv6 would handle 99.9% of all usages.

> IPv4 doesn't seem to need the user to tell it to use udp, it seems to default 
> to 
> that. Any reason not to make the default switch to udp6 if dns returns *only* 
> an 
> AAAA record? (I'm not even going to go into the insanity that could result 
> from 
> trying to do both if DNS returns both an A and an AAAA address.)

The problem is that the name lookup happens after the choice of
transport at the moment.

> It looks like a bit of code around line 380 of snmplib/snmp_transport.c might 
> be 
> the ticket to accomplishing that.

In a way this is true, but that is realy kind of a wart on the system,
it exists solely to make sure that some old special cases keep working.

I think bug #1801793 is another angle on this very same problem, and the
solution I proposed there with multiple default transports probably
would work on this problem as well.

> > actually does make snmptrapd receive and print the trap.
> 
> Yup, technically it does work... just in a very user-un-friendly way.

Well, yes, there is a potential to increase the usability.

/MF


-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Net-snmp-coders mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/net-snmp-coders

Reply via email to