Connectivity issues and packet inspection

[Steve Bertrand]

Phil Pennock wrote:
> On 2008-06-18 at 21:25 -0400, Steve Bertrand wrote:
>> The lack of /64 is all I needed to hear in order for it to make sense. 
>> I'll look up some documentation on how to change it.
> 
> netsh takes '?' as a command in various places and if you don't supply
> enough parameters to a command, it offers help.
> 
> It had been a while since I last touched Windows but I just, uhm,
> borrowed the use of my wife's XP box.

...I use XP every day, but I don't *use* it.

It's 0430 here now, and I am required back at work at 0730, so it's time 
for bed. I just caught your message as I was printing some docs, and 
without getting into detail, wanted to state:

> Last time I looked at this stuff on XP, I was trying to set up a static
> tunnel through a NAT to my FreeBSD box and I don't think I ever did get
> it going.  Shocker, that.

My opinion with XP *personally* at this point, is that since I 
configured the static IP/gw originally via the 'command line' months 
ago, it hasn't failed me.

The only problem I have ever had began today, and it wasn't an XP 
problem, it apparently turned out to be an intercontinental transient 
routing issue beyond XPs control.

I just picked up on the 'hit the router for a redirect' issue as I was 
troubleshooting something outside of the scope of my local /64 (which 
has been made apparent to me, is likely not a /64 at all).

>>> Vista support of IPv6 seems much better than XP,
>>> so I'm told.

There is a Vista box that I could throw my XP box at right now, and I 
have full intentions on testing many aspects of IPv6 capabilities and 
standards with.

In particular, there are a few RFCs, and more specifically continuously 
updated I-Ds that I have serious interest in that I'd like to follow 
along with to keep current edge. My problem is not enough me, and not 
enough resources to have more than one me.

> XP sometimes seems to decide that there are no IPv6 addresses on the
> LAN, even though the Apple Airport Extreme clearly is advertising the
> routes.  It can go months not messing up, then mess up a couple of times
> a week.   *sigh*  The only fix I've gotten is, I think, to disable IPv6
> on the interface and re-enable it again; interface repair doesn't work,
> disabling and re-enabling the interface doesn't work.  It's cmdline
> time and using netsh.
> 
> So for "normal" users for whom the command-line is not an option, XP
> isn't a reliably useful client over IPv6.

v6 IMHO, from someone who from time to time tries to keep control of an 
ISP helpdesk, is a piece of nothingness that a user will appreciate as 
much as they do their toaster. If it works, great. If not, they get a 
new one.

If my staff have to explain to a user 'right-click on x connection', 
then forget it. It's too much work. From my perspective, I need to 
implement it and use it on my box to get a feel for it. My ultimate 
objective is to obtain the knowledge and experience that the supposed 
benefits of 4861 and 4862 have, so 'dummies' can get online with no 
interaction required.

I don't care shim6, 464 etc. I try to follow the RRG, but it's just too 
far above my head. I hope a standard is created that is seamless in 
regards that my North American ADSL client can attach to any BAS, on any 
Version of IP Protocol, grab an IP from RADIUS, and just go with the flow.

...the work that I'm trying to do on my theory however, is based around 
the src addr selection, M & O bits etc, etc that will impact all of this.

We are much farther with 'auto-discovery', but we need to clear up a few 
small areas... God willing, now that we have prefixes and rtradv in 
place, someone who has the resources to 'play' needs to develop a spec 
to simply pass but one name server address, and a true *plug and play* 
specification will be born...

...tired, overspoken, off to bed. Thanks all...

Steve