Extra routes causing server to stop talking

Question

I have 16 Sun M3000 servers divided into 2 systems of 8.
System A is on 192.168.217.0 255.255.255.0 /etc/defaultrouter = 192.168.217.1
System B is on 192.168.218.0 255.255.255.0 /etc/defaultrouter = 192.168.218.1
Routing between the networks is done via cisco asa5520 Firewalls.
I am using 2 NICs on each server in a single group with one configured as a standby:
/etc/hostname.bge0 = server1 group serv1grp up
/etc/hostname.bge0 = group serv1grp standby up
The problem I have is that under normal conditions, when I type netstat -r I see:
217.0 network:
Routing Table: IPv4
Destination Gateway Flags Ref Use Interface
default 192.168.217.1 UG 1 101
192.168.217.0 192.168.217.11 U 1 120 bge0
224.0.0.0 192.168.217.11 U 1 0 bge0
127.0.0.1 127.0.0.1 UH 3 65 lo0
218.0 network:
Routing Table: IPv4
Destination Gateway Flags Ref Use Interface
default 192.168.218.1 UG 1 173
192.168.218.0 192.168.218.12 U 1 80 bge0
224.0.0.0 192.168.218.12 U 1 0 bge0
127.0.0.1 127.0.0.1 UH 3 65 lo0
Unfortunately every once in a while (8 hours or so) the servers on the 218.0 network quit talking to the servers on the 217.0 network. No ping, etc.. I can get from a server on the 217.0 network, through the firewalls and switches to a serrver on the 218.0 network. All of the 218.0 servers can talk to each other but not back over to 217.0 at this point. They can see 218.1
I have noticed "extra routes" in the route table on these servers. I do not have the exact extras, but something simliar to this:
218.0 network:
Routing Table: IPv4
Destination Gateway Flags Ref Use Interface
default 192.168.218.1 UG 1 173
192.168.218.0 192.168.218.12 U 1 80 bge0
192.168.218.0 192.168.218.34 U 1 80 bge0
192168.218.4 192.168.218.12 U 1 80 bge0
224.0.0.0 192.168.218.12 U 1 0 bge0
127.0.0.1 127.0.0.1 UH 3 65 lo0
I understand that by having the /etc/defaultrouter file, routing/forwarding is disabled by default and I have not re-enabled any of this.
It seems as though the server is learning invalid routes. What I want to do is basically force it to always have only theoriginal 4 and not learn any new ones.
Anyone have any ideas?

Answer 1

Dont seem to find any difference between other servers...i chnaed the timeout settings to 
"ServicesPipeTimeout" to 90000..This seems to have minimized the issue...but i am not sure for long it';s gonna work...
Initially the issue came to timeouts on 30000ms so i had to increase it to 60000
first Issue chnage :1: A timeout (30000 milliseconds) was reached
..Settings changed to 60000.it worked for a couple of month and then started to throw errors agaiin :
Second :A
timeout (60000 milliseconds) was reached 
Now i have chnaged the value to 90000,though it doesnot show any recommended value for the reg keys for "ServicesPipeTimeout" and any fixes for
services getting timeout..
A timeout (60000 milliseconds) was reached while waiting for a transaction response from the CryptSvc service.
A timeout (60000 milliseconds) was reached while waiting for a transaction response from the Dnscache service.
A timeout (60000 milliseconds) was reached while waiting for a transaction response from the LanmanWorkstation service.
A timeout (60000 milliseconds) was reached while waiting for a transaction response from the WinRM service.
A timeout (60000 milliseconds) was reached while waiting for a transaction response from the NlaSvc service.