routing table not updating

[Guy Humphreys]

Hello all,

We have got a problem with an old 10.20 box. Yes, I know, I know we are planning on moving to something a little bit more supported but you know how it is.

The problem is sometimes a WAN link will go down and it's ISDN backup route will kick in automatically. all well and good. The problem comes when the ISDN is no longer needed and the main WAN link comes back. the networking side of things all functions nicely but the routing table on the 10.20 box doesn't update it still thinks the remote end is connected via the ISDN and will try and direct traffic down this now none existent link.

How can I tell what the timeout period for new routes to be discovered is and alter it to say 5 minutes?

Thanks a lot
Guy

[Mugilvannan]

Paste your netstat -rn output here. Based on routing setting only we can control routing. I hope there is no explicit setting for timewait.

hth.

[Guy Humphreys]

I have already deleted the offending routes myself manually. So a netstat -rn will only show a working config not the problematic one.

but here it is:

/ $netstat -rn
Routing tables
Destination Gateway Flags Refs Use Interface Pmtu PmtuTime
127.0.0.1 127.0.0.1 UH 0 96 lo0 4608
172.18.1.215 127.0.0.1 UH 3 2027012 lo0 4608
172.18.4.15 172.18.1.252 UGHD 0 93185 lan0 1500
172.18.4.16 172.18.1.252 UGHD 0 62947 lan0 1500
172.18.4.17 172.18.1.252 UGHD 0 5842829 lan0 1500
172.18.4.50 172.18.1.252 UGHD 0 56776 lan0 1500
172.18.4.58 172.18.1.252 UGHD 0 27 lan0 1500
172.18.4.63 172.18.1.252 UGHD 0 596 lan0 1500
172.18.4.65 172.18.1.252 UGHD 0 18204 lan0 1500
172.18.4.70 172.18.1.252 UGHD 0 5791 lan0 1500
172.18.4.71 172.18.1.252 UGHD 0 39058 lan0 1500
172.18.4.72 172.18.1.252 UGHD 0 314 lan0 1500
172.18.4.188 172.18.1.252 UGHD 0 11089 lan0 1500
172.18.4.190 172.18.1.252 UGHD 0 490911 lan0 1500
172.18.4.191 172.18.1.252 UGHD 1 622073 lan0 1500
172.18.4.195 172.18.1.252 UGHD 1 482993 lan0 1500
172.18.4.196 172.18.1.252 UGHD 0 778914 lan0 1500
172.18.4.197 172.18.1.252 UGHD 1 625389 lan0 1500
172.18.4.200 172.18.1.252 UGHD 0 51642 lan0 1500
172.18.4.201 172.18.1.252 UGHD 0 38438 lan0 1500
172.18.4.202 172.18.1.252 UGHD 0 38627 lan0 1500
172.18.4.203 172.18.1.252 UGHD 0 13283 lan0 1500
172.18.4.204 172.18.1.252 UGHD 0 33822 lan0 1500
172.18.4.205 172.18.1.252 UGHD 0 9682 lan0 1500
172.18.4.206 172.18.1.252 UGHD 0 22881 lan0 1500
172.18.4.207 172.18.1.252 UGHD 0 45877 lan0 1500
172.18.4.208 172.18.1.252 UGHD 0 11819 lan0 1500
172.18.4.209 172.18.1.252 UGHD 0 4534 lan0 1500
172.18.4.210 172.18.1.252 UGHD 0 47321 lan0 1500
172.18.4.211 172.18.1.252 UGHD 0 31831 lan0 1500
172.18.4.212 172.18.1.252 UGHD 0 11993 lan0 1500
172.18.4.213 172.18.1.252 UGHD 0 44511 lan0 1500
172.18.4.214 172.18.1.252 UGHD 0 127783 lan0 1500
172.18.4.215 172.18.1.252 UGHD 0 375413 lan0 1500
172.18.4.216 172.18.1.252 UGHD 0 31 lan0 1500
172.18.4.217 172.18.1.252 UGHD 0 21 lan0 1500
172.18.4.235 172.18.1.252 UGHD 1 468246 lan0 1500
172.18.4.236 172.18.1.252 UGHD 0 292631 lan0 1500
172.18.4.237 172.18.1.252 UGHD 1 704720 lan0 1500
172.18.25.186 172.18.1.252 UGHD 0 322509 lan0 1500
default 172.18.1.253 UG 6533872635 lan0 1500
172.18 172.18.1.215 U 29170928748 lan0 1500

Cheers
Guy

[Mugilvannan]

Routing table is followed using /etc/rc.config/netconf file when you do reboot or shutdown. Pls change settings of route in that file accordingly.

[Muthukumar_5]

Are you sure that,

172.18 172.18.1.215 U 29170928748 lan0 1500

this is from netstat -rn output??

Plz give your ISDN and WAN Link IP-Address

hth.

[Muthukumar_5]

I hope your problem is because of dead gateway.

Pls check this,

# ndd /dev/ip ip_ire_gw_probe_interval
180000
#
#
# ndd -h ip_ire_gw_probe_interval

ip_ire_gw_probe_interval:

Controls the probe interval for Dead Gateway Detection.
IP periodically probes active and dead gateways.
ip_ire_gw_probe_interval controls the frequency of probing.
With retries, the maximum time to detect a dead gateway is
ip_ire_gw_probe_interval + 10000 milliseconds. Maximum time
to detect that a dead gateway has come back to life is
ip_ire_gw_probe_interval. [15000,- ]
Default: 180000 (3 minutes)

# ndd /dev/ip ip_ire_gw_probe
1

Change as,
# ndd set /dev/ip ip_ire_gw_probe 1
# ndd set /dev/ip ip_ire_gw_probe_interval 300000

I think it will work 5 minutes now.

hth.

[Guy Humphreys]

yep 172.18.0.0 is our local subnet range

the main WAN link goes through the default gateway 172.18.1.253 so is not present in the routing table

the ISDN link, when it is active is something like 192.206.75.1 going via the default gateway again. I can't recreate it as our main link is fine at the moment.

the problem is that once this 192.206 route is in the routing table it doesn't get deleted ever even though it is a dynamic route created by an ICMP re-direct from the main default router

Cheers
Guy

[Guy Humphreys]

Muthukumar, I can't use ndd as I am on 10.20 only got nettune I am afraid.

[Mugilvannan]

Pls change the network tunable parameter of ip_ire_gw_probe_interval to 5 minutes and change it into /etc/rc.config.d/nddconf
for future usage as,

TRANSPORT_NAME[2]=ip
NDD_NAME[2]=ip_ire_gw_probe_interval
NDD_VALUE[2]=300000

save the file.

hth.

[Muthukumar_5]

Can you see nettune man page and list available variables. I hope nettune -l will result all. There may be a variable related to ip_ire_gw_probe_interval. ( :( 10.20 No box is with me)

hth.

[Guy Humphreys]

OK, here is a print out of all nettune parameters:

arp_killcomplete = 1200 default = 1200 min = 60 max = 3600 units = seconds
arp_killincomplete = 600 default = 600 min = 30 max = 3600 units = seconds
arp_unicast = 300 default = 300 min = 60 max = 3600 units = seconds
arp_rebroadcast = 60 default = 60 min = 30 max = 3600 units = seconds
icmp_mask_agent = 0 default = 0 min = 0 max = 1
ip_check_subnet_addr = 0 default = 1 min = 0 max = 1
ip_defaultttl = 255 default = 255 min = 0 max = 255 units = hops
ip_forwarding = 1 default = 1 min = 0 max = 1
ip_intrqmax = 50 default = 50 min = 10 max = 1000 units = entries
pmtu_defaulttime = 20 default = 20 min = 10 max = 32768
tcp_localsubnets = 1 default = 1 min = 0 max = 1
tcp_receive = 32768 default = 32768 min = 256 max = 262144 units = bytes
tcp_send = 32768 default = 32768 min = 256 max = 262144 units = bytes
udp_receive = 9360 default = 9360 min = 256 max = 262144 units = bytes
tcp_defaultttl = 64 default = 64 min = 0 max = 255 units = hops
tcp_keepstart = 600 default = 7200 min = 8 max = 12000 units = seconds
tcp_keepfreq = 60 default = 75 min = 5 max = 2000 units = seconds
tcp_keepstop = 300 default = 600 min = 10 max = 4000 units = seconds
tcp_maxretrans = 12 default = 12 min = 4 max = 12
tcp_urgent_data_ptr = 0 default = 0 min = 0 max = 1
udp_cksum = 1 default = 1 min = 0 max = 1
udp_defaultttl = 64 default = 64 min = 0 max = 255 units = hops
udp_newbcastenable = 1 default = 1 min = 0 max = 1
udp_pmtu = 0 default = 0 min = 0 max = 1
tcp_pmtu = 1 default = 1 min = 0 max = 1
tcp_random_seq = 0 default = 0 min = 0 max = 2
so_qlimit_max = 4096 default = 4096 min = 1 max = 8192
sb_max = 262144 default = 262144 min = 10240 max = 4294967295
hp_syn_protect = 0 default = 0 min = 0 max = 1
so_qlimit_min = 500 default = 500 min = 0 max = 8192
high_port_enable = 0 default = 0 min = 0 max = 1
high_port_max = 65535 default = 65535 min = 49153 max = 65535
ip_forward_directed_broadcasts = 1 default = 1 min = 0 max = 1
tcp_fin_wait_timer = 0 default = 0 min = 0 max = 16383 units = seconds

As you can see , there are a lot less here than in ndd. none of them seem like they are close to what you have in 11i.

You are lucky not having any 10.20s let me tell you! :)

cheers
Guy

[Bob_Vance]

You should consider giving more points to the replies.
When someone takes the time to think about your problem and then make suggestions, you should appreciate that and give more than 0 or 1 point.

(0 for this is OK, though :>)

[Guy Humphreys]

Bob, thanks for the interest, I am just going on the ITRC guidelines though:

o N/A: The answer was simply a point of clarification to my original question

o 1-3: The answer didn't really help answer my question, but thanks for your assistance!

o 4- 7: The answer helped with a portion of my question, but I still need some additional help!

o 8-10: The answer has solved my problem completely! Now I'm a happy camper!

All of the answers given were either not 10.20 related and so although valid, did not solve any part of my problem or were questions asking for more details and so were awarded 0 points. Now I know 10.20 is unsupported so I am just 'throwing this out there' so to speak to see if anyone has had a similar experience.

Besides if I gave too many points away too easily there would be no challenge! :)

Thanks to all that replied, the irony is as soon as I move to 11i I bet this problem goes away on its own!

Guy