Richard wrote regarding 'Re: [SLE] Subnet 169.254.0.0' on Thu, Aug 19 at 08:21:

On Thursday 19 August 2004 07:30 am, Jerry Feldman wrote:

...

I noticed last night while trying to configure my SuSE 9.1 laptop for MIT, that the routing table had en entry for 169.254.0.0. It is coming from the network scripts, ifup-route. It shouldn't hurt, but I was wondering why this was done, or is it something that someone at SuSE set up. I haven't seen any documentation on this either.

Jerry, if you do a whois 169.254.0.0 you will see it eventually goes to some server called BLACKHOLE. As I understand it, that's what it is. When xp fails to get the proper IP from the DHCP server, it gets one of the 169.254 numbers and nothing works. The strange thing is, as soon as Linux fires up on the same machine I get the correct local IP. I'm sure there is some longwinded answer somewhere but the short one is it's a dummy IP.

Now some guru can step in and tell us the REAL STORY!

169.254.0.0/255.255.0.0 is the netblock used by autoconfigured DHCP machines. If you've ever taken a windows box and turned it on without a DHCP server anywhere, it will autoconfigure itself to use an address from that range. It'll then send out a broadcast to see fi anyone else replies on that address, and pick another from that range if the one it chose was taken. And so on. That way, you could have a hub, some wires, and no knowledge of how to configure a network, but still get a couple of home machines to talk to each other. Most DHCP clients now will use that so they can interoperate nicely. --Danny

Quoting Paul Ikanza :

Anyone using postfix mail notications to cell phone/pager Any Ideas? Paul.

Not quite sure what you are asking. Most cellphone service providers will forward e-mail sent to a certain address to your cellphone's SMS, e.g., 9999999999@vtext.com for Verizon, 9999999999@mobile.att.com for AT&T, 9999999999 is your ten digit phone number (i.e., including area code). The MTA doesn't matter, Postfix, Sendmail, Exim, Qmail, whatever. This is in the United States, YCMV (Your Country May Vary). There are SMS gateway packages, don't know how they work. What they require in terms of hardware and telco services. HTH, Jeffrey

Does your phone have email notification? I will be doing this when my new phone arrives.. Neal -----Original Message----- From: Paul Ikanza [mailto:PIkanza@newvision.co.ug] Sent: Thursday, August 19, 2004 6:47 AM To: 'SuSE List' Subject: [SLE] Notifications Anyone using postfix mail notications to cell phone/pager Any Ideas? Paul. -- Check the headers for your unsubscription address For additional commands send e-mail to suse-linux-e-help@suse.com Also check the archives at http://lists.suse.com Please read the FAQs: suse-linux-e-faq@suse.com

Jerry wrote regarding 'Re: [SLE] Subnet 169.254.0.0' on Thu, Aug 19 at 09:40:

On Thu, 19 Aug 2004 08:48:25 -0500 Danny Sauer wrote:

...

169.254.0.0/255.255.0.0 is the netblock used by autoconfigured DHCP machines. If you've ever taken a windows box and turned it on without a DHCP server anywhere, it will autoconfigure itself to use an address from that range. It'll then send out a broadcast to see fi anyone else replies on that address, and pick another from that range if the one it chose was taken. And so on. That way, you could have a hub, some wires, and no knowledge of how to configure a network, but still get a couple of home machines to talk to each other. Most DHCP clients now will use that so they can interoperate nicely.

Yes, but I don't understand why it is configured on 9.1, even when I have used only a static address.

Say you've got a windows machine at home, too, and that you haven't set up a DHCP server. You just went out, bought a network card, and plugged it into your windows machine. Then you hooked both of your machines up to a hub. Now your windows machine will assign itself an address in that range by itself, without you having to do anythign else. If that address is configured on the SuSE box, then the two computers will be able to see each other with no other configuration. It doesn't hurt anything, and it possibly makes it easier to interoperate with the windows box. I'm guessing that's why it was done. Or maybe you're running the dhcpd client for some reason, even with the static IP? Then again, I'm just guessing at this point. My 9.1 boxes all use DHCP, so I don't see that here. :) --Danny

On Thursday 19 August 2004 08:48 am, Danny Sauer wrote:

169.254.0.0/255.255.0.0 is the netblock used by autoconfigured DHCP machines. If you've ever taken a windows box and turned it on without a DHCP server anywhere, it will autoconfigure itself to use an address from that range.  It'll then send out a broadcast to see fi anyone else replies on that address, and pick another from that range if the one it chose was taken.  And so on.  That way, you could have a hub, some wires, and no knowledge of how to configure a network, but still get a couple of home machines to talk to each other.  Most DHCP clients now will use that so they can interoperate nicely.

--Danny

I understand your answer, although I'm not real certain why things are done that way. I have a Dell 1150 laptop with XP & SuSE9.1 on it and it has a real problem getting an IP from my Belkin wireless router when running XP. It usually picks up the blackhole IP and will not get a useable IP. But then sometimes after a restart it will pick up the correct IP and work fine until the next boot. Linux on the same laptop gets the correct IP every time. For me that's enough of an excuse to delete XP now. On the other hand, I have a friend with an HP laptop using XP and an ethernet connection going into a linksys router/hub and that machine has the same problem but it will hook up properly about 80% of the time. It will work 100% if I plug it into my linux router, indicating some kind of router problem, maybe. Is there a reliable way to insure the client gets the proper IP. Or is there a way to force the client to get the IP? I realize this is really a windows question but I ask it here since 99% of my time is spent using SuSE and I would like to know why Linux is so much easier to connect to a DHCP server. Richard

What's serving DHCP, the linksys router/hub or the linux router - or are these on seperate networks?   They are two separate networks in two different places. One uses the

On Thursday 19 August 2004 02:47 pm, Danny Sauer wrote: linksys router/hub which connects to the internet via cable modem and provides masquerading for the local net. It is the dhcp server for the local net which is all windows stuff 98 and XP. The 98 machines request and get a consistent IPs assigned and work fine. It's the XP machine that has a hard time getting the IP and often defaults to the blackhole number. That same XP machine when plugged into my separate linux router never fails to get its IP assigned by the dhcp server. At the second location I have the wired stuff going into a linux router/server and it feeds into my wireless (Belkin) route which then goes to the cable modem. I know it's a little strange but I did it that way cause I had VoIP /dmz set up on the linux machine to handle my phone system independently. The wireless was added to primarily take care of laptops. Again, linux based wireless gets the IP without fail while the same laptop running XP will often come up with blackhole number and only rebooting will let it connect with the proper IP - sometimes. I have another laptop running win98 which always gets the proper IP.

Basically, a DHCP client sends out a broadcast asking for an IP, the DHCP server responds with a broadcast that the machine should get, and then the machine sends another message to the DHCP server (Ack!  my Address! - or something like that :)).

So, perhaps the broadcasts aren't being passed around properly, or perhaps you've just got a bad cable - does the setup work better under Linux v/s windows when using the exact same ethernet cable in the exact same port on the exact same hub? Yes! I put linux on the hp laptop and it gets the IP right every time! Using same everything only the router is different and the linux hookup is also without a problem!

I wonder what diagnostic tool will tell me what is happening with the windows stuff. Richard

On Monday 23 August 2004 05:15 pm, Danny Sauer wrote:

Personally, I'd fire up ethereal just before turning the windows machine on, every time.  Look for the DHCP packets, and see what the heck's going on. You might get similar help from ntop, but I think ethereal will do what you want there.

--Danny, slow to respond, so it seems

I'll give that a try, Thanks. ra