[Gllug] Anyone having broadband (probably PlusNet) problems today?

[Russell Howe]

On Wed, Mar 23, 2005 at 01:35:29PM +0000, Richard Jones wrote:
> 
> Woke up today to find that my PlusNet internet connection behaving
> very oddly indeed.  It's not that it doesn't work, just that certain
> types of connections hang.

This reeks of MTU breakage, where either ICMP time-exceeded messages are
being ignored, being blocked or not being sent at all.

Try pinging a remote host with ever-increasing ping sizes. You want to
be making sure you are setting the DF (don't fragment) bit on the IP
datagrams when you do so.

With the ping program I have in Debian, this is done like this:

ping -M do -s <size>

e.g.:

PING www.bbc.net.uk (212.58.224.126) 1462(1490) bytes of data.
1470 bytes from www26.thdo.bbc.co.uk (212.58.224.126): icmp_seq=1 ttl=250 time=81.7 ms

Note the two numbers on the 1st line, 1462 and 1490. 1462 is the size of
the ping packet, 1490 is the size of the IP datagram containing the ping
packet. Here, this shows that the smallest MTU between me and
www26.thdo.bbc.co.uk is at least 1490 (unless some piece of hardware is
fragmenting an IP datagram which has the DF bit set, which it Should Not
Do(tm) AFAIK).

If you are running PPPoA on a DSL link, your MTU is likely 1500. If you
are running PPPoE, your MTU is lower (1492?) due to the overhead of
PPPoE on top of ethernet's 1500 byte MTU.

I won't do my whole DSL-sucks-and-there-is-no-decent-CPE-hardware rant
here again :)

-- 
Russell Howe       | Why be just another cog in the machine,
rhowe at siksai.co.uk | when you can be the spanner in the works?
-- 
Gllug mailing list  -  Gllug at gllug.org.uk
http://lists.gllug.org.uk/mailman/listinfo/gllug