No route to host ??

[Bejoy C Alias]

We have two servers connected to the internet through leasedline
which are used as the web/mail and dns servers . if i do a traceroute to
www.hclinsys.com from our web/mail server

[root@server root]# traceroute -n www.hclinsys.com
traceroute to www.hclinsys.com (203.90.70.241), 30 hops max, 38 byte packets
1 210.212.163.177 4.062 ms 3.634 ms 4.062 ms
2 210.212.71.18 28.573 ms 28.645 ms 28.419 ms
3 61.0.228.70 28.774 ms 29.282 ms 28.265 ms
4 218.248.255.6 52.681 ms 52.545 ms 51.532 ms
5 218.248.250.81 80.301 ms 80.158 ms 81.355 ms
6 218.248.250.73 79.708 ms 80.245 ms 80.202 ms
7 218.248.249.42 80.998 ms 80.727 ms 80.695 ms
8 * * *
9 218.248.251.90 65.349 ms 66.292 ms 65.737 ms
10 198.32.202.17 358.817 ms 365.857 ms 364.032 ms
11 203.90.67.181 355.572 ms 359.632 ms 355.232 ms
12 203.90.66.137 364.572 ms 361.020 ms 846.468 ms
13 203.90.66.138 1436.081 ms 2894.837 ms 3401.283 ms
14 203.90.66.137 2366.757 ms 1921.396 ms 2292.976 ms
15 203.90.66.138 1329.353 ms 1473.352 ms 2536.595 ms
16 203.90.66.137 1306.155 ms 582.747 ms 1435.093 ms
17 203.90.66.138 596.953 ms 1649.929 ms 2858.607 ms
18 203.90.66.137 1151.034 ms
In the output , after the 12 th hop it is going in loop between ip xxx.137 and xxx.138
but from the second server the traceroute is proper which is also
using the same leased line link.
[root@ns2 root]# traceroute -n www.hclinsys.com
traceroute to www.hclinsys.com (203.90.70.241), 30 hops max, 38 byte packets
1 210.212.163.177 3.671 ms 4.038 ms 3.725 ms
2 210.212.71.18 877.108 ms 1093.133 ms 571.859 ms
3 61.0.228.70 2173.317 ms 394.703 ms 919.037 ms
4 218.248.255.6 578.504 ms 67.130 ms 1358.458 ms
5 218.248.250.81 224.710 ms 1368.238 ms 2303.700 ms
6 218.248.250.73 1232.395 ms 389.839 ms 752.307 ms
7 218.248.249.42 743.280 ms 940.543 ms 1103.702 ms
8 * * *
9 218.248.251.90 2241.985 ms 1910.239 ms 2039.044 ms
10 198.32.202.17 2041.005 ms 1503.732 ms 1152.707 ms
11 203.90.67.181 1132.641 ms 1122.030 ms 1621.174 ms
12 203.90.66.137 2416.331 ms * 3358.409 ms
13 203.76.132.161 3088.667 ms 1278.685 ms 608.595 ms
14 203.90.70.161 953.908 ms 1344.621 ms 2324.226 ms
15 203.90.70.234 3094.415 ms 3065.216 ms 1780.932 ms

If i do a ping
from the first server it is giving the error as "time to live exceeded"
and this error is from the 12th hop (xxx.137).
Our first server is the mail server and we are unable to send mails
to the domain hclinsys.com as it is giving the "no route to host" error .
from the second server i am able to send mails to this domain.

What will be the pblm ??

[Ralph Grothe · Accepted solution]

Can you ask the admin of the 12th hop to check their routing table?
Why is it forwarding packets from your 1st server to the wrong next hop 203.90.66.138 that doesn't know where to route them to,
whereas it correctly knows the next hop for packets from your 2nd server?

[Siljumon S]

Hi Bejoy,

90% this problem is at the router where it is where it is going into a loop. As you said you can reach the destination from the second server, just compare your default ttl settings of both the servers. If any difference try increasing the ttl by putting an entry in the file

/etc/sysctl.conf and put the one line like below
ip_default_ttl = "ttlvalue"


here "ttlvalue" 64 is default Try increasing this number.
You may require a restart after this change.

Warm Regards

Silju

[Bejoy C Alias]

This pblm is solved .
I didn't do anything at my servers . I think the admin of the 12th hop may came to know about the issue and made the modifications at their routers .
Thanks to all ...