From owner-freebsd-questions Fri Jun 4 8: 2:12 1999 Delivered-To: freebsd-questions@freebsd.org Received: from cc942873-a.ewndsr1.nj.home.com (cc942873-a.ewndsr1.nj.home.com [24.2.89.207]) by hub.freebsd.org (Postfix) with ESMTP id 8F62B14F52 for ; Fri, 4 Jun 1999 08:02:09 -0700 (PDT) (envelope-from cjc@cc942873-a.ewndsr1.nj.home.com) Received: (from cjc@localhost) by cc942873-a.ewndsr1.nj.home.com (8.9.3/8.8.8) id KAA03219; Fri, 4 Jun 1999 10:53:35 -0400 (EDT) (envelope-from cjc) From: "Crist J. Clark" Message-Id: <199906041453.KAA03219@cc942873-a.ewndsr1.nj.home.com> Subject: Re: Question about arp entry in /var/log/messages In-Reply-To: <199906040728.AAA29672@cup44ux.cup.hp.com> from Ken Lui at "Jun 4, 99 00:28:33 am" To: klui@cup.hp.com (Ken Lui) Date: Fri, 4 Jun 1999 10:53:35 -0400 (EDT) Cc: dwhite@resnet.uoregon.edu, questions@FreeBSD.ORG Reply-To: cjclark@home.com X-Mailer: ELM [version 2.4ME+ PL40 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Ken Lui wrote, > >From dwhite@resnet.uoregon.edu Thu Jun 3 23:52:39 PDT 1999 > > Doug, > > I must thank you for having the patience in helping me get to the > root of my problems. > > > > I've run two tcpdumps per interface and with the exception of some > > > items at the beginning and the numbers after the timestamp, they're > > > the same. Looks like both interfaces are seeing packets on net 10 > > > and net 15. Here are the entries that show up under both dumps after > > > the following entry ends up in /var/log/messages: > > > Jun 1 21:14:05 black /kernel: arp: 10.0.0.1 is on lo0 but got reply from > > > 00:80:c8:fd:88:0d on ed1 > > > > _lo0_? Hm! Can I see ifconfig -a, please? Try to keep the whole message > > around since I'm trying to keep track of this. lo0 should be there. What is strange is that ed1 is hearing things for 10.0.0.1. > Yeah, I thought about that but did so after I sent the last reply. > > ifconfig -a > ed1: flags=8843 mtu 1500 > inet 15.75.136.174 netmask 0xfffff800 broadcast 15.75.143.255 > ether 00:80:c8:fd:90:ae > ed2: flags=8843 mtu 1500 > inet 10.0.0.1 netmask 0xffffff00 broadcast 10.0.0.255 > ether 00:80:c8:fd:88:0d > lo0: flags=8049 mtu 16384 > inet 127.0.0.1 netmask 0xff000000 > > > Do you have proxy arp turned on in the cisco? > > I checked it and it doesn't have proxy arp capability. I did notice > that its netmask is incorrect for its IP. IT did this for me so I > never questioned it. Changing its netmask didn't eliminate those > entries. The frustrating thing is these messages are have a certain > delay before they show up in /var/log/messages. > > Odd how after I've ping'ed some of these IPs, I get the following > with netstat -nr > > Routing tables > > Internet: > Destination Gateway Flags Refs Use Netif Expire > default 15.75.136.169 UGSc 4 13 ed1 > 10/24 link#2 UC 0 0 ed2 > 10.0.0.1 0:80:c8:fd:88:d UHLW 0 28 lo0 > 10.0.0.2 link#2 UHLW 0 2 ed2 > 10.0.0.3 link#2 UHLW 0 2 ed2 > 10.0.0.4 8:0:7:6f:1d:fe UHLW 0 15 ed2 828 > 15.75.136/21 link#1 UC 0 0 ed1 > 15.75.136.169 0:40:f9:13:69:d5 UHLW 5 4 ed1 764 > 15.75.136.174 0:80:c8:fd:90:ae UHLW 0 24 lo0 > 127.0.0.1 127.0.0.1 UH 1 4 lo0 > > So somehow, lo0 is receiving messages which ed1 and ed2 should be > receiving. That is correct behavior. The addresses to the loopback are the addresses of the machine's own interfaces. That netstat looks almost exactly like it should. As a point of reference, the machine I am on now has two interfaces, here are the netstat -rn entries, Internet: Destination Gateway Flags Refs Use Netif Expire default aaa.bbb.ccc.1 UGSc 4 18440 fxp0 127.0.0.1 127.0.0.1 UH 1 2509 lo0 192.168 link#2 UC 0 0 192.168.0.1 0:a0:c9:22:93:d0 UHLW 0 2 lo0 192.168.0.2 0:aa:0:a5:af:91 UHLW 2 8392 fxp1 59 192.168.0.255 ff:ff:ff:ff:ff:ff UHLWb 1 352 fxp1 aaa.bbb.ccc link#1 UC 0 0 aaa.bbb.ccc.1 0:e0:1e:3e:40:0 UHLW 3 0 fxp0 778 aaa.bbb.ccc.2 0:aa:0:6f:d7:28 UHLW 3 37287 fxp0 607 aaa.bbb.ccc.5 0:10:5a:18:44:67 UHLW 0 0 fxp0 1132 aaa.bbb.ccc.6 0:a0:c9:23:b:ad UHLW 0 2990 fxp0 1077 aaa.bbb.ccc.10 8:0:69:7:40:16 UHLW 4 331675 fxp0 483 aaa.bbb.ccc.102 0:90:27:13:9e:cf UHLW 0 4 fxp0 1147 aaa.bbb.ccc.111 0:a0:c9:27:53:d6 UHLW 0 2 fxp0 1118 aaa.bbb.ccc.148 0:a0:c9:9d:f1:c UHLW 0 2 fxp0 925 aaa.bbb.ccc.154 0:a0:c9:27:f:5e UHLW 0 4 fxp0 1002 aaa.bbb.ccc.194 0:90:27:10:a:8b UHLW 0 2 fxp0 1150 aaa.bbb.ccc.204 0:90:27:13:ae:7e UHLW 0 130918 lo0 aaa.bbb.ccc.206 0:60:97:8a:4d:68 UHLW 0 2 fxp0 1069 aaa.bbb.ccc.214 0:aa:0:bb:2b:5e UHLW 0 2 fxp0 926 aaa.bbb.ccc.255 ff:ff:ff:ff:ff:ff UHLWb 2 2234 fxp0 Note the loopback addresses for the machine's own interfaces. Also note the broadcast addresses... *shrug* dunno where yours are. > Maybe a dumb and naive question but... when I was running Linux I > didn't have to run gated or routed. Should I be running it? You shouldn't need it for simple, static routing like this. > > > 21:14:05.461124 arp who-has green.tmpest1.org tell black.tmpest1.org > > > 21:14:05.461600 arp reply green.tmpest1.org is-at 8:0:7:6f:1d:fe > > > > Is this the proper ether addr for green? > > Yes, I've confirmed that its ethernet address is correct. > > > > The beginning of ed1 (net 15) has the following when I first > > > establish a connection to my router (start of dump): > > > 21:10:28.449996 ce573230.cup.hp.com.iad3 > 15.75.12.3.domain: 1784+ (37) > > > 21:10:29.390619 ce573230.cup.hp.com.1033 > 15.75.12.3.domain: 23899+ (43) > > > > Lots of DNS lookups but no responses. > > > > > While ed2 (net 10) has the following (start of dump): > > > 21:11:48.500727 ce573230.cup.hp.com.1040 > 15.75.12.3.domain: 1785+ (60) > > > 21:11:48.572032 ce573230.cup.hp.com.1041 > 15.75.12.3.domain: 6263+ (43) > > > > ?? What is that stuff going that way? When you run tcpdump for this stuff, use the -n option. That way, we can see which interfaces are being used rather than the name of a machine that might have multiple interfaces. > > Let me clarify this. The interfaces are listed next to the IPs they're > > assigned, if I'm getting you right. > > Yes. For instance, ce573230 should be on ed1, black should be on ed2. > However, green is multi-homed with 1 ethernet interface so it really > rests on both ed1's and ed2's network; but the confusing thing for > me is they both share one wire. Whoa. Did you just say ed1 and ed2 are on the same wire? You _will_ have trouble if two interfaces on one machine share a wire. There is no really good reason I am aware of to ever have that situation. -- Crist J. Clark cjclark@home.com To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message