On Mon, 17 Feb 2003, at 16:43 [=GMT-0500], IAccounts wrote:

> > >Feb 16 18:35:06 voo /kernel: arplookup failed: host is
> > >not on local network
> > >Feb 16 18:35:06 voo /kernel: arpresolve: can't allocate llinfo for
> > >
> I had this back in the summer time, and it was due to having an IP
> address aliased on one of my nics in a block that was not on the local
> subnet.

Thanks for your eply. Alas, this is not the case. The machine has 3 IP
numbers from one half class C from my ISP. And a separate full class C
(199.a.b/24) is routed to it. This all works fine. The IP number the
arplookup complaints are about is from another server, which I have
collocated elsewhere.

For a while I thought I had caused the problem by giving two of the
three IPs my ISP gave me the same reverse DNS names as two on my other
(old) server, which at the time I was considering to take down after
two months or so, but didn't. I described this a few months back here.
Basically reverse DNS now sees the same name for (names and numbers

OLD                                     NEW
apple.exter.net              apple.exter.net
pear.joyrider.nl             pear.joyrider.nl
                                        u.com2.us (main

The normal DNS only gives out for apple.exter.net. The old
thing. So if the new machine does a lookup for the IP of its own
apple.exter.net, it gets, and not the IP on its own NIC, Then it starts looking for and sending these
arp complaints.

This is what I thought. But I think it isn't relevant. For two

1. Why not similar problems for pear.joyrider.nl?
2. Why still problems when I change the DNS for apple.exter.net and
give it two IPs in the zone for exter.net?
3. Since reverse DNS is not only a mess with me (because I have to
deal with sys admins of my ISP to get it changed...), but all over the
net, I would think this problem would occur a lot more, if wrong
reverse DNS caused it.

> Check the block the NIC's bound IP is on, then verify that any blocks that
> the aliases are in are reachable from that machine.

Everything works fine, only at times syslog is very busy. By the way,
it stopped as miraculously as it started in the mean time. Which makes
me believe it is a problem caused by some router. I would like to
understand it though...

[01] All ideas are vintage not new or perfect.

To Unsubscribe: send mail to [EMAIL PROTECTED]
with "unsubscribe freebsd-questions" in the body of the message

Reply via email to