I have 2 IP Lease Circuits ,terminating on 7204 - NPE-G1 ,
utilization for Router 4-5 % . Running Eigrp in same AS.
Problem Description:
The 2 links have latency L1- 150 ms and L2- 200 ms, we have ip cef
flow switching enabled on both sides. Since Delay is factor for eigrp
metrics, L1 is preffered,
eigrp uses "Delay" as part of its composite metric, but this delay is
fixed per interface (shown in "show int" and does not depend on the
RTT).
but often situation becomes bad, Link1 is
utilized to about 80% while Link2 is utilized only about 30% , though
after sometimes, the situation becomes normal , and they both might
run at 40% / 30% etc..
Can you send your eigrp and interface config and send "show ip route
x.x.x.x" and "show ip cef x.x.x.x internal" for a sample route going
over the interfaces.
To me this currently sounds like normal behaviour.. since we hash on L3
src/dest, a low number of flows can cause unequal distribution.
We cannot run per packet load balancing since that has made our vioce
issues worse.
Any ideas to tackle this situation?
Multilink PPP?
oli
2 messages in net.nether.puck.cisco-nsp[c-nsp] Re: EIGRP load balancing.
