Carp seems completely broken on 8.2-RC2 and 8.2-PRERELEASE
freebsd at psconsult.nl
Mon Jan 17 13:25:20 UTC 2011
On Mon, Jan 17, 2011 at 01:05:31PM +0100, Daniel Hartmeier wrote:
> On Sun, Jan 16, 2011 at 01:41:22PM +0100, Paul Schenkeveld wrote:
> > There is an ARP request which is replied to by the carp master (test).
> > the ping to the carp address does not even appear on the sis4 interface
> > of test1.
> Everything looks fine, except for the fact that the ping (echo request)
> doesn't get to test1's sis4.
> Are you sure the problem isn't with the switch? Have you tried resetting
> it? Or replacing it with another one (where you could check the MAC
> address table, etc.)?
The switch has been power-cycled, no change. Only 3 ports are
wired, to test1, test2 and test3. I'm not in the office right now, can
replace the switch tonight, but read on...
> You'd get this behavior if the switch had learned carp4's virtual MAC
> address (00:00:5e:00:01:68) on another port. You're not using vhid 104
> (:68 in the virtual MAC) on other ports of that switch, are you?
test3 has no carp nor vrrp so vhid 104 is not in use anywhere else.
Tcpdump shows only carp (vrrp) packets from test1 one per second.
sis3 of test1 and test2 are connected by a cross-cable. IP addresses
are 10.3.0.1/24 (carp3, vhid 103, test1 is master, test2 is backup),
10.3.0.2/24 for sis3 on test1 and 10.3.0.3 for sis3 on test2.
On test1 I can ping 10.3.0.1 (which test1 is carp master for), from
test2 I can't ping 10.3.0.1. A tcpdump on sis3 on test1 shows ARP
request and reply, but no icmp echo-request. The arp entry on test2
test2 # arp 10.3.0.1
? (10.3.0.1) at 00:00:5e:00:01:67 on sis3 expires in 800 seconds [ethernet]
On test2 I can ping 10.3.0.2 and 10.4.0.2 (the addresses on sis3 and sis4
of test1) and see the normal arp-request/arp-reply/icmp-echoreq/
icmp-echoreply sequence using tcpdump.
More information about the freebsd-stable