[Date Prev][Date Next] [Thread Prev][Thread Next]
[Thread Index]
[Date Index]
[Author Index]
Re: no kickstart ?
- From: Eric Hagberg morganstanley com
- To: taroon-beta-list redhat com
- Subject: Re: no kickstart ?
- Date: Thu, 31 Jul 2003 20:03:36 -0400 (EDT)
The bugzilla is here:
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=97892
I'll have to try out that patch tomorrow, though I suspect this will fix
things.
We have our routers configured to forward dhcp packets appropriately so
that we don't have to run dhcp servers on every network.
Thanks,
-Eric
On Fri, 1 Aug 2003, Florian La Roche wrote:
> On Thu, Jul 31, 2003 at 05:36:56PM -0400, Eric Hagberg morganstanley com wrote:
> > I've also reported a problem where DHCP fails during install (it uses
> > pump) but works just fine when the server boots into the installed copy of
> > RHEL3 (which uses dhclient).
> >
> > >From snooping, it appears to be due to the TTL on pump's discovery packets
> > being set to 0, while dhclient has it set to something like 16. Unless
> > your dhcp server is on the same network and you don't need your dhcp
> > requests forwarded by your routers, the installer will fail, as packets w/
> > TTL=0 are discarded at the first router they hit.
>
> So your subnets have a dhcp-server running that is forwarding the
> requests to another server?
>
> >
> > I've opened an issue on this, but it doesn't look like it's getting fixed
> > soon.
>
>
> I have checked the "pump" bugzilla list and didn't find anything. Quick
> inspection of the pump source code look like this should be a one-line
> change to the pump sources by changing "hopcount" in pump to non-0:
>
> --- ./dhcp.c.lr 2003-08-01 00:05:06.000000000 +0200
> +++ ./dhcp.c 2003-08-01 00:05:20.000000000 +0200
> @@ -595,7 +595,7 @@
> for (i = 0; i < IFHWADDRLEN; i++)
> breq->xid ^= breq->hwaddr[i] << (8 * (i % 4));
>
> - breq->hopcount = 0;
> + breq->hopcount = 16;
> breq->secs = 0;
>
> initVendorCodes(breq);
>
> Dhcpcd seems to set this to 64 now. Any reason to not apply the above
> change?
>
> The failure to work with Windows 2000 could be something completely
> different. Can you check running "pump" with the ttl patch? Anyone
> who has debugged the Win2000 server about the root cause?
>
> greetings,
>
> Florian La Roche
>
[Date Prev][Date Next] [Thread Prev][Thread Next]
[Thread Index]
[Date Index]
[Author Index]