Hello,
Using RB1100AHX2, ROS 6.27.
One direct connected IP and a /29 subnet.
No NAT, No firewall, No QoS, No extra routes, No BGP or similiar.
Problem
From any linux machine on the /29 subnet behind the router, traceroute behaves weird.
Meaning:
~# traceroute google.com
traceroute to google.com (82.77.159.212), 30 hops max, 60 byte packets
1 noc.mydomain.org (XX.XXX.XXX.XX) 0.131 ms 0.067 ms 0.108 ms
2 static-X-X-XXX-X.myisp.org (X.X.XXX.X) 2.048 ms 2.292 ms 2.195 ms
3 * * *
4 * * *
5 * * *
6 * * *
7 * * *
8 * * *
9 * * *
10 * * *
11 * * *
12 * * *
13 * *^C
After the second HOP, there is a small delay resolving (I guess) the 3rd and final HOP.
Adding -I (ICMP) or -T (TCP SYN) to traceroute results in instant response:
~# traceroute -I google.com
traceroute to google.com (82.77.159.251), 30 hops max, 60 byte packets
1 noc.mydomain.org (XX.XXX.XXX.XX) 0.119 ms 0.137 ms 0.138 ms
2 static-X-X-XXX-X.myisp.org (X.X.XXX.X) 1.083 ms 1.204 ms 1.330 ms
3 cache.google.com (82.77.159.251) 15.436 ms 15.455 ms 15.458 ms
~# traceroute -T google.com
traceroute to google.com (82.77.159.251), 30 hops max, 60 byte packets
1 noc.mydomain.org (XX.XXX.XXX.XX) 0.119 ms 0.137 ms 0.139 ms
2 static-X-X-XXX-X.myisp.org (X.X.XXX.X) 1.083 ms 1.204 ms 1.331 ms
3 cache.google.com (82.77.159.251) 15.436 ms 15.455 ms 15.437 ms
Using MTR result in instant trace also.
Using traceroute domain.org without any switches, results in timeout right at the last HOP.
Very random and very rarely, a simple traceroute domain.org will finish without timeout.
Also, traceroute to some external destinations ALWAYS finish without timeout.
Everything else works just right, traffic is ok, bandwidth is ok.
Question
Given that I am a greenhorn in RouterOS, maybe I did something wrong configuring the router?
But if I did, wouldn’t that mean all destination should timeout or work? Not only some fail and some work?
Is there any setting in RouterOS that can cause this kind of problem?
Thank you.