I'm looking of your OSPF config and I got some questions for you...

1- is there any 0.0.0.0 area define somewhere ? if not, change your area 
0.0.0.1 to 0.0.0.0 for test... if you're having a 0.0.0.0 area, make 
sure this one is define in your config...

2- can you try to place your CARP interface in passive mode : aka to not 
send ospf packet on these interfaces...
example:
            interface carp1 {
                passive
        }

3- in your global config... try to define:
    no redistribute connected
    no redistribute static
    redistribute x.x.x.x/xx

here is some old comment that I left on my config to remind me some old 
good time...

##
## Do not redistribute all connected routes
## automatically, only redistribute routes from
## participating interfaces in OSPF or that we
## explicitly want to redistribute. First rule is
## evaluated first.
##
##
## Hit my head on the wall on this one, interfaces
## participating in OSPF will be redistributed in the RIB
## even if you add a "no redistribute" statement. Only add
## the interfaces that you want participating into an area
## to the area definitions. the "passive" flag makes an interface
## part of that area but disables the broadcast of OSPF packets
## on that interface. Interfaces that are not defined anywhere
## do not participate in OSPF at all.

-
Marcus

Joe Warren-Meeks wrote:
quoted text
> On Wed, May 28, 2008 at 04:53:32PM +0100, Joe Warren-Meeks wrote:
>   
>> On Wed, May 28, 2008 at 11:37:52AM -0400, Marc-Andre Jutras wrote:
>>     
>>> I had mostly the same issue with my HP Proliant DL320 G5  with obsd 4.1 
>>> and  4.2....
>>>
>>> got some packet lost on the interface without any error on my cisco 
>>> switch -> making my ospf flip and re-calculate route too every 10 sec...
>>>
>>> I change all the network cards on this server and it fix my issue...  
>>> looks like there some issue with the BGE / Broadcom drivers...
>>>
>>> as a test, try to run a 'mtr'  or a high rate ping ( ping -c 10000 -i 
>>> 0.01 ) on your interface... make sure you're not getting any packet lost 
>>> for at least 10 min...
>>>       
>> I'm seeing a load of working packets, followed by a load of "No route to
>> host" and back again.
>>
>> Think I'll try swapping to a different card.
>>     
>
> I have swapped out the bge card for an em card and am getting exactly
> the same thing. Periodically, the routes go screwy as follows:
>
> # route get 195.24.231.41
>    route to: 41                <============= See here
> destination: 41
>   interface: em1
>  if address: 3
>       flags: <UP,HOST,DONE,LLINFO,CLONED>
>      use  hopcount       mtu    expire
>   158767         0         0       658 
> # route get 195.24.231.41  
>    route to: 195.24.231.41     <============ And back again
> destination: 195.24.231.41
>   interface: em1
>  if address: 195.24.231.3
>       flags: <UP,HOST,DONE,LLINFO,CLONED>
>      use  hopcount       mtu    expire
>   242792         0         0       557 
> #
>
> Any clues?
>
>  -- joe.
>
> Why do you have a hairdryer? You don't even have hair.