topic RE: Issue with exact 50% icmp ping loss in ExtremeSwitching (EXOS/Switch Engine)

Issue with exact 50% icmp ping loss

Raj_Kam — Mon, 29 Feb 2016 19:35:00 GMT

Hi Gens,

facing issue with Extreme switches X460 .
As when i am pinging the other node onm vip from the extreme l3 switch

getting the ping response with every alternate ping loss like below
64 bytes from x.x.x.x: icmp_seq=0 ttl=64 time=1.195 ms
64 bytes from x.x.x.x: icmp_seq=2 ttl=64 time=0.963 ms
64 bytes from x.x.x.x: icmp_seq=4 ttl=64 time=1.237 ms
64 bytes from x.x.x.x: icmp_seq=6 ttl=64 time=1.030 ms

and the seq are even for one switch and with other it is odd like below.

64 bytes from x.x.x.x: icmp_seq=1 ttl=64 time=1.105 ms
64 bytes from x.x.x.x: icmp_seq=3 ttl=64 time=0.863 ms
64 bytes from x.x.x.x: icmp_seq=5 ttl=64 time=1.155 ms
64 bytes from x.x.x.x: icmp_seq=7 ttl=64 time=1.475ms

is this the issue with the switch or it is the expected behavior.

help me how to troubleshoot it .

Regards
Raj

RE: Issue with exact 50% icmp ping loss

Stephen_Elliot1 — Mon, 29 Feb 2016 19:40:00 GMT

this sort of thing is often incorrectly configured LAGs - are you using port sharing? is it working correctly at both ends?

RE: Issue with exact 50% icmp ping loss

Henrique — Mon, 29 Feb 2016 19:49:00 GMT

Hi Raj, would be great to have some network diagram to understand the traffic directions and what's configured between them (LAG, MLAG, VRRP, etc).

RE: Issue with exact 50% icmp ping loss

Raj_Kam — Mon, 29 Feb 2016 20:01:00 GMT

wait let me share that 

RE: Issue with exact 50% icmp ping loss

Raj_Kam — Mon, 29 Feb 2016 20:25:00 GMT

Hi

we have two routes towards VIP
#oa 10.196.x.x/32 10.196.a.b 5 UG-D---um--f- route_1_1 21d:9h:52m:40s
#oa 10.196.x.x/32 10.196.a.c 5 UG-D---um---- route_2_1 21d:9h:52m:40s

both have same metric of 5.

and if i ping from the other node then the output is same as alternate ping goes miss .
so from both the side i am facing the same issue.

ospf is working between the node and the two switches

Regards
Raj

RE: Issue with exact 50% icmp ping loss

Raj_Kam — Mon, 29 Feb 2016 20:31:00 GMT

and traceroute to VIP from SWITCH

# traceroute 10.196.x.xtraceroute to 10.196.X.X, 30 hops max
1 * * *
2 10.196.x.x 3 ms * 2 ms

--- Packet Response/Error Flags ---
(*) No response, (!N) ICMP network unreachable, (!H) ICMP host unreachable,
(!P) ICMP protocol unreachable, (!F) ICMP fragmentation needed,
(!S) ICMP source route failed, (!u) Transmit error, network unreachable,
(!f) Transmit error, fragmentation needed, (!t) General transmit error
# traceroute 10.196.x.x
traceroute to 10.196.X.X, 30 hops max
1 * * *
2 *
2 10.196.x.x 2 ms *

--- Packet Response/Error Flags ---
(*) No response, (!N) ICMP network unreachable, (!H) ICMP host unreachable,
(!P) ICMP protocol unreachable, (!F) ICMP fragmentation needed,
(!S) ICMP source route failed, (!u) Transmit error, network unreachable,
(!f) Transmit error, fragmentation needed, (!t) General transmit error

RE: Issue with exact 50% icmp ping loss

Henrique — Mon, 29 Feb 2016 20:41:00 GMT

Even with those 2 OSPF routes, the switch is only using the route through route_1_1. You can confirm that based on the flag "f" that means "provided to FIB".

The second route is active but not being used.

It seems to be related to asymmetric routing.

I would check from point A to B and B to A if that communication is using the same path.

RE: Issue with exact 50% icmp ping loss

Raj_Kam — Tue, 01 Mar 2016 00:32:00 GMT

need to check the routing ?

RE: Issue with exact 50% icmp ping loss

Henrique — Tue, 01 Mar 2016 01:30:00 GMT

Correct, I would check the routing on each hop as well as LAG/VRRP configuration.