We have one Cisco Router 6506. When I ping the local interface of the router, it responds always within 1ms. But every 2 minutes, the Router latency is at ~800ms for nearly one second. This repeats every 2 minutes.
What could be the reason for this periodically delays?
Patrick
Ping to the router are treated as a low priority task.
Traffic thru the box is treated different than traffic to the box.
Ping to a device on the other side of thw switch; it is a much better test of the switch latency.
This is not how network latency is usually defined. What matters is the delay that packets passing through the router are subjected to. This test is not measuring that delay.
The 6500 (in general) forwards traffic in /hardware/ and the CPU is used for management tasks only. When you ping the router you are pinging the CPU. This is not necessarily a problem. The delay of 200ms or so is however quit large and I would want to try to get some idea of what was going on. We see delays of a few 10s of ms when we ping our 4500s (SE 4, 5).
This is /no indication whatsoever/ that you voice traffic is at risk.
"What is the reason?" This could be hard to figure out. Start by checking router CPU.
What happens on /your/ network every 2 mins?
Merv schrieb:
Sombody told me that already. But ...
... this doesn't help in this case. When I do VoIP to an other Vlan over this router, VoIP is ugly in the same second.
snipped-for-privacy@hotmail.com schrieb:
I pinged other devices from other vlans. The delay is the same every 2 minutes. Hosts in the same vlan are not affected.
Actually, it does. VoIP was the reason why I found this bug with our router. As this is our only realtime-application, we didn't notice it earlier with other TCP/UDP-Applications.
I don't know as I am not the network admin, I report only errors to him ;-) I did a check with ethereal but I could not find anything exceptional (broadcast, Multicast).
If pings thru the switch are delayed every 2 minutes then something is amiss.
If you are not the network administrator then it is going to be hard to help you as responders would need some output from the router in question in order to be able to assist. Perhaps you can get you netadmin to post required info
Required output for troubleshooting:
sh version
sh process cpu
sh ip cef
sh ip arp sum
sh running-config
sh int status
sh int
Are there any ugly spikes in CPU or "CPUHOG" messages in syslog?
I can't think of any normal periodic activity that happens every *two* minutes.
Need more clues. Does the traffic level also spike at that time?
"Patrick Cervicek" wrote in message news:e47kn6$4q$ snipped-for-privacy@news.BelWue.DE...
As it was told before, it'd be hard to troubleshoot since you are not Network Admin. I would recommend to connect to the Console on the switch, and see if you see any messages at the time when you have your switch delays. Somehting MUST be in the log - errors, warnings, activity reported, etc. It could be your starting point.
Good luck,
Mike
HeadsetAdapter.com schrieb:
Today I informed the network admin. He will handle this problem and contact this group within the next days. As far as we saw, there were nothing exceptional in the logs. Do we need a certain debug-/log-flag set?
Patrick
Phillip Remaker schrieb:
We didn't see any 'CPU' messages.
No. Nothing in syslog and nothing in mrtg.
I did more tests.
Pings to other vlans (Esslingen,same location) are not affected - but one. We have a BSD-VPN-Router connected to this interface/vlan, which is connected to an other router (B) at an other location. All Hosts after B are delayed after 2 minutes for one second
Ping_Host Router_A BSD-VPN-Router VPN_Router_B $any_HOST_at_B
The interesting part is, that pings to the BSD-Router and Router B are never affected.
Here you have an mtr-graph from Ping_Host (look at Wrst)
17:32:09 2006I want to do VoIP with Location_B, but you see that everything is delayed after 2 minutes (at the same time, when Route_A's Latency is high , $any_HOST_at_B is too)
We will check that tomorrow, Patrick
We had a statement "evaluate blablalist" (Reflexive-ACL) in our incoming-Accesslist. This Reflexive-ACL is very big.
We put the "permit ip host $VPN_Router_B host $BSD-VPN !Allow VPN-Router_B to us"
*before* that statement in our Incoming-ACL, and now we have very good ping results to location B. VoIP works!
We think, that (efficient) Hardware-Routing was impossible as the Hardware had always to lookup the dynamic ACL before passing traffic through.
Thank you for your hints!
Patrick
Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.