I am one of those others who is possibly encountering this also, however I am using a basic Dachstein setup.
Will let you know if it works. Steve On Thu, 27 Jun 2002 22:24:11 -0500 "Ethan Galstad" <[EMAIL PROTECTED]> wrote: > I sent a message to the leaf-devel list a few days back about this > issue, but didn't get any replies. Any developer willing to > recompile dhclient with this TTL hack (one line of code)? > > I tried compiling ISC's dhcp-2.0pl5 on a RedHat 5.2 box today - > compiled okay and executed okay on my LRP box, but I get > "send_packet" errors when trying to send DHCPDISCOVERS to eth0. The > stock > dhclient-2.0pl5 in the Dachstein 1.0.2 image doesn't give these > errors, so > its driving me mad. :-( Since its driving me crazy, I was hoping > someone > who actually knew what they're doing (not me) could recompile. I > know > there must be others out there who have encountered this issue. > > > > ------- Forwarded message follows ------- > From: Ethan Galstad <[EMAIL PROTECTED]> > To: [EMAIL PROTECTED] > Subject: DHCP Client TTL Problem - AT&T Broadband > Date sent: Sat, 22 Jun 2002 19:08:40 -0500 > > There is a problem with the ISC's DHCP client (including all LRP > version) with the TTL being hardcoded to 16. This started giving me > (and > others) problems last Monday when AT&T broadband "upgraded" their > systems > in the St. Paul/Minneapolis area. > > Turns out that now the DHCP server is (no joke) 18 hops away from me. > > Since the DHCP client sets the DHCPDISCOVER packet TTL to 16, the > message never gets to the server. The "fix" is to hack the ISCP DHCP > source code and use a larger value (i.e. 128) in the TTL field. > AFAIK, > the Microsoft DHCP client uses a value of 128 in the TTL field. > > The message below was sent by Paul Dokas where he outlines how to do > this. > > Can someone make an updated dhclient package with this change > incorporated? FWIW, I'm running the Dachstein 1.0.2 distribution. > My home network is useless without it. :-( > > PS: I found a newsgroup message outlining the same problem and fix, > so I know there are others out there who are having similar problems. > > Here's the Google URL: > > http://groups.google.com/groups?selm=c91381b2.0206210920.269f9e1b%40po > stin > g.google.com > > > ------- Forwarded message follows ------- > > [snip] > > Found the problem. The ISC DHCP client sets the TTL on the DHCP > request to 16, which is apparently too low for AT&T's network. This > would explain why it works for some people and not other on an > apparently random basis. > > > I edited line 159 of common/packet.c in the ISC DHCP source code. It > used > to be: > > ip.ip_ttl = 16; > > and I changed it to be: > > ip.ip_ttl = 128; > > > And, now I get a lease almost immediately after running dhclient. > > > So, I guess that AT&T's DHCP servers are just fine (I don't need to > set any strange options in dhclient.conf). The "problem" is that the > Internet, or at least AT&T's broadband network, has grown larger than > what > the source code assumes to be a *big* network. > > > I'll send in a bug report to ISC first thing in the morning. > > > Paul > ------- End of forwarded message ------- > > > > > Ethan Galstad, > Nagios Developer > --- > Email: [EMAIL PROTECTED] > Website: http://www.nagios.org > > > > ------------------------------------------------------- > This sf.net email is sponsored by:ThinkGeek > Bringing you mounds of caffeinated joy. > http://thinkgeek.com/sf > ------------------------------------------------------------------------ > leaf-user mailing list: [EMAIL PROTECTED] > https://lists.sourceforge.net/lists/listinfo/leaf-user > SR FAQ: http://leaf-project.org/pub/doc/docmanager/docid_1891.html ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Bringing you mounds of caffeinated joy. http://thinkgeek.com/sf ------------------------------------------------------------------------ leaf-user mailing list: [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/leaf-user SR FAQ: http://leaf-project.org/pub/doc/docmanager/docid_1891.html
