On Mon, Nov 30, 2009 at 09:31:19PM +0000, Stuart Henderson wrote:
> On 2009-11-30, stan <st...@panix.com> wrote:
> > Sorry, I am still confused here. What I have is a pair of machines, each
> > machine has 3 physical interfaces. On each machine one is for the "inside"
> > network, one is for the "outside" network, and one is for phsync. The
> > inside network is a single subnet, and does not need to see OSPF routing,
> > as all of it's machines have a static default route to the CARP'd "inside"
> > interface.
>
> so use interface carpXX { passive } for this one...
>
> > Both the "inside" and "outside" interfaces on both machines have
> > an equiv. CARP interface. So, there are 3 outside IP addresses. the CARP
> > address, and an individual address for the outside interface on each
> > machine.
>
> ...and use the real interfaces for these, not the carp ones.
> You shouldn't need carp on the outside interfaces.
>
> > What can't happen is to have the machines both advertise their
> > real physical interface addresses as duplicate routes to the inside
> > network, right?
>
> In 4.6 and earlier, only the carp master advertises the inside network.
>
> In -current, both master and backup announce it, master with a low metric
> so it's preferred, backup with a high metric. so the route isn't normally
> used but it isn't totally lost when the routers failover.
>
First of all, thanks to everyone for working so hard to educate me. I am
trying to learn here.
Now, I have turned off the external carp interface, and things still work,
but when I tried chnaging the ospfd.conf file I killed rotuing to the
internal network. Here is what I see at the moment:
ifconfig shows:
carp1: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
lladdr 00:00:5e:00:01:02
priority: 0
carp: MASTER carpdev em0 vhid 2 advbase 1 advskew 100
groups: carp
inet 170.85.106.143 netmask 0xffffff80 broadcast 170.85.106.255
inet6 fe80::200:5eff:fe00:102%carp1 prefixlen 64 scopeid 0
and it shows the internal interface as up. But:
s...@phfw2:stan$ ospfctl show interfaces
Interface Address State HelloTimer Linkstate Uptime ncac
em0 170.85.106.145/25 DOWN - active 00:00:00 0 0
bge0 10.209.142.153/25 BCKUP 00:00:06 active 06:20:40 2 2
If I change the intrenal interface to the carp interface I get:
r...@phfw2:etc# ospfctl show interfaces
Interface Address State HelloTimer Linkstate Uptime nc ac
carp1 170.85.106.143/25 DOWN - master 00:00:00 0 0
bge0 10.209.142.153/25 BCKUP 00:00:04 active 06:40:45 2 2
This does not look correct. Is it?
This is no a 4.6 set of machines, BTW:
with ospfd.conf files that look like this:
area 0.0.0.120 {
interface bge0 {
auth-type none
}
interface carp1 {
passive
auth-type none
}
}
Here is what I see:
On the machine with carp in MASTER:
r...@phfw2:etc# ospfctl show interfaces
Interface Address State HelloTimer Linkstate Uptime nc ac
carp1 170.85.106.143/25 DOWN - master 00:00:00 0 0
bge0 10.209.142.153/25 BCKUP 00:00:08 active 06:47:21 2 2
On the nachine with carp in BACKUP
r...@phfw1:etc# ospfctl show interfaces
Interface Address State HelloTimer Linkstate Uptime nc ac
carp1 170.85.106.143/25 DOWN - backup 00:00:00 0 0
bge0 10.209.142.152/25 OTHER 00:00:06 active 06:46:33 2 2
This does not give me confidence that this is working corectly.
Am I mistaken?
--
A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in e-mail?
