From: John Gibson (johngibson1541@yahoo.com)
Date: Thu Apr 05 2007 - 04:53:19 ART
Sorry about the confusion. I read the ipv6 rfc 2460
for the first tonight. I see they say routers in
the path MUST NOT fragment any ipv6 packet.
At the same rfc , it also says every ipv6 router in
the path MUST support MTU equal or larger than
1280. And in the same rfc, it also points out if
we don't want to deal with fragments, simply implement
the end hosts to send packets smaller than 1280 bytes.
It turned out, they basically say if you can't do
MTU discovery, just send packets smaller than 1280
bytes.
phew thats lots of tricks going into ipv6. they
learned
these good tricks from the experience of ipv4 ?
John
--- fonesurj <dwinkworth@wi.rr.com> wrote:
> Not sure what you are asking here.
>
> In an IPv6 network, the routers will not fragment
> the packet. So you're
> ping will fail if it is too large. In that case,
> you should get a
> datagram-too-large ICMPv6 message back from the
> router that can't pass the
> traffic.
>
> So if you have dot1q tunneling "on" and for some
> reason that reduces your
> path MTU, the ping will fail if the ping is larger
> than allowed. It won't
> know why it failed or what kind of link caused the
> failure, but it will
> fail.
>
>
>
> ----- Original Message -----
> From: <johngibson1541@yahoo.com>
> To: <ccielab@groupstudy.com>
> Sent: Wednesday, April 04, 2007 10:10 AM
> Subject: Re: IPv6: routers don't fragment any
> packet. End hosts all MUST
> have path MTU discovery ?
>
>
> > If we use ping to ping an IPv6 host across our
> lab, if we set the
> > data size to test if various packet size can go
> through our pad,
> > our routers will fragment the data according to
> the MTU discovery ?
> >
> > If there is a misconfiguration in our switches in
> our pad (such as
> > dot1q tunneling) , the ping can not detect the
> misconfiguration ?
> > Because the path MTU discovery will detect the
> packet size that
> > can go through our pad lossless ? How do we know
> if we have MTU
> > misconfiguration in IPv6 then ?
> >
> >
>
This archive was generated by hypermail 2.1.4 : Tue May 01 2007 - 08:28:34 ART