Re: Routing problem
On Tue, 2006-02-14 at 12:14, jb701@uku.co.uk wrote:
> On Mon, 2006-02-13 at 13:43, jb701@uku.co.uk wrote:
> Your network looks like this:
>
> .------. .-----------. .-----------.
> | A | | B | | C |
> | .2.2 +---+ .2.1 .1.2 +---+ .1.1 .0.6 +--- .0.*
> `------' `-----------' `-----------'
>
> When you attempt to send a packet from .2.2 to .1.1, how does system A
> know where to send it?
>
> Assuming such a packet reaches .1.1, how does system C know where to
> send the reply?
>
> You may need some routes. Perhaps these will allow A and C to talk.
>
> On A# route add default gw 192.168.2.1
> On C# route add -net 192.168.2.0/24 gw 192.168.1.2
>
> You have four or more hosts on three or more network segments. I'd
> strongly recommend an introductory networking course or book. With
> a firm understanding of the principles there's no limit to what you
> can achieve.
>
> --Mike Bird
>
> Thanks Mike. I have routes set up as you describe, and when I look at the
> routing cache on the debian box [...1.2 and ...2.1] using route -Cn, packets
> from 192.168.1.1 trying to get to 192.168.2.2, and vice versa, appear in the
> list. This is from using ping from ..2.2 to ..1.1 and the other way round.
>
> That means they are arriving in the debian machine, but it seems not to be
> doing much with them. The flag shown is "i", but man route doesn't explain
> the meaning of that.
Use the source Joe. :-)
The "i" flag appears to be associated with "RTF_IRTT" which
is something to do with calculation of initial round trip
time. Probably not relevant to your problem.
At this point I'd wind up "ethereal" - or maybe just "tethereal"
for looking at pings - and see how far the packets and/or replies
were travelling.
--Mike Bird
Reply to: