Weird temporary problem with quagga routing

[Reiner030]

Hi,

we have 2 pfsense pairs for 2 buildings… the "old" side works correctly with 2.0.1 but in new site with "2.1-BETA1 (amd64) built on Wed Feb 27 04:47:52 EST 2013" I have again and again temporary routing problems...

The firewall slave lose his internally routes  (and very rare time also the master):

I have e.g. our icinga/nagios monitoring on network of side 1 and we check all firewalls and network equipment  of building 2, too:

Icinga/192.168.1.9  ->  192.168.1.1/2/3 [pfsense-pair jws1] 192.168.6.1/2/3  <=== Transfer ===> 192.168.6.11/12/13 [pfsense-pair zws8] 192.168.8.1/2/3 -> switch admin network

If the slave 192.168.6.12 is not reachable by icinga I got this routing:

[2.1-BETA1][root@fw2.zws8.local]/root(6): route -n get 192.168.1.9
  route to: 192.168.1.9
destination: default
      mask: default
    gateway: xx.xx.176.254  <=public IP GW
  interface: lagg0_vlan7        <= WAN interface
      flags: <up,gateway,done,static>recvpipe  sendpipe  ssthresh  rtt,msec    mtu        weight    expire
      0        0        0        0      1500        1        0

but OSPF rooting seems allright:
[2.1-BETA1][root@fw2.zws8.local]/root(7): vtysh -e "show ip route 192.168.1.9"
Routing entry for 192.168.1.0/24
  Known via "ospf", distance 110, metric 20
  Last update 05:03:21 ago
    192.168.6.2, via lagg0_vlan6
    192.168.6.3, via lagg0_vlan6

So my problem is… what take this buggy route alive ?

It can't be an BETA problem because I hade this also on our 2.0.1-er firewalls but very rarely there:

[2.0.1-RELEASE][root@fw1.jws1.local]/root(2): route -n get 192.168.66.11
  route to: 192.168.66.11
destination: default
      mask: default
    gateway: xx.xx.176.1  <= public GW
  interface: lagg0_vlan2    <= WAN interface
      flags: <up,gateway,done,static>recvpipe  sendpipe  ssthresh  rtt,msec    mtu        weight    expire
      0        0        0        0      1500        1        0

Interesting thing what helps in this cases:
[2.0.1-RELEASE][root@fw1.jws1.local]/root(2): /etc/rc.reload_interfaces

[2.0.1-RELEASE][root@fw1.jws1.local]/root(3): route -n get 192.168.6.11
  route to: 192.168.6.11
destination: 192.168.6.11
    gateway: 192.168.6.12
  interface: lagg1_vlan6
      flags: <up,gateway,host,done,proto1>recvpipe  sendpipe  ssthresh  rtt,msec    mtu        weight    expire
      0        0        0        0      1500        1        0

If OSPF is shutdown all works long time without problems.

I found an OpenOSPF Problem/Fix and tested it out:
http://forum.pfsense.org/index.php/topic,39995.0.html
links to his blog:
http://ouliakk.blogspot.com/2011/08/using-openospfd-with-freebsd-78.html

The Fix / Workaround

The concept is simple: create an IP alias where the network overlaps the existing IP/network.
In our example, 192.168.11.0/24 is used to exchange OSPF information. Create an alias of 192.168.10.1/23. That way when the 192.168.11.0/24 route gets deleted, the systems will be accessible to each other over the 192.168.10.0/23 route.

but this solution didn't help in this case :(

Andy ideas please?

Bests

Reiner</up,gateway,host,done,proto1></up,gateway,done,static></up,gateway,done,static>