Networking problems on finishing LFS

Bill's LFS Login lfsbill at nospam.dot
Thu Feb 12 16:35:05 PST 2004


On Fri, 13 Feb 2004, Ashley Oliver wrote:

> Okay, just been into the Linux system to try some of the suggestions, now
> back in Win2K...
>
> > It sounds like there's a problem with your network route.
> > What does the output of the "route" command look like?
>
> 'route' produces this: (I'll list the two lines vertically as there's too
> many columns to fit properly horizontally)
>
> Entry 1:
>
> Destination:	192.168.1.0
> Gateway:		*
> Genmask:		255.255.255.0
> Flags:		U
> Metric:		0
> Ref:			0
> Use:			0
> Interface:		eth0
>
> Entry 2:
>
> Destination:	default
> Gateway:		192.168.1.2 (My router's IP)
> Genmask:		0.0.0.0
> Flags:		UG
> Metric:		1
> Ref:			0
> Use:			0
> Interface:		eth0
>

One thing to keep in mind here. *Unless* you have routing setup to
handle it, your pings will never get to the default gateway. Default is
taken *only* if no prior route can be found, IIRC. With the netmask and
addy's I recall, only eth0 will get hit with the above.

To further isolate, do ifconfig down eth0 (and you don't need default
when it's down either) and try your pings. Maybe that will be the clue
you need?

Also, the pings could be arriving, but getting routed to the wrong eth
when your machine responds?

Initial testing should be as simple as possible for problem isolation.

> > Can you ping this machine from another one?
>
> Nope. My other machines all run Windows, and they all report a timeout when
> trying to ping the linux machine.

Possibly because of the route prob above?

><snip>

> My card is a PCI one, with some realtek chipset. Can't remember the model
> now... RTL31...something, but when I found out what it was I saw Linux
> actually included a driver for the exact model, so that's compiled into the
> kernel. Seems to be okay, since Linux detected it fine, and brings up the
> interface with no problems.

If it's the 8139too driver, my experience is that driver may be flaky as
hell. Check you logs for messages along the lines of "too many
interrupts" or "too much work" or something similar. If you see any of
those, post and I'll pass on what I know that works arond it (doesn't
fix it, but works around it).

>
> I might rebuild all the network-related files from scratch tommorow, and
> check the things I selected to be put in the kernel, and recompile that too,
> just to make absolutely sure everything is as it is supposed to be... but
> I'm pretty sure it is.
>
> One little thing I noticed... even after a few failed ping attempts, the
> statistics of the eth0 interface still show 0 sent 0 recieved. Shouldn't
> there at least be some sent?

What's eth1 showing? Do the counts go up or does it also stay unchenged
on one or both counts?

>
> Off to bed now, will try more tommorow. Thanks for suggestions so far.
>
> Ashley
>
>

-- 
NOTE: I'm on a new ISP, if I'm in your address book ...
Bill Maltby
lfsbillATearthlinkDOTnet
Fix line above & use it to mail me direct.



More information about the blfs-support mailing list