Description

Frequent next-hop changes in the network for a large number of routes can lead to high heap utilization on the Compact Forwarding Engine Board (CFEB). This article describes how to check route flaps/network changes.

Symptoms

How to troubleshoot high heap utilization on CFEB.

Solution


High heap utilization can be caused by simple exhaustion of the memory, such as too large of a routing table or too many route flaps/changes in the network. Frequent next-hop changes in the network for a large number of routes can lead to high heap utilization. It can also indicate a memory leak or a hardware problem on the CFEB.

Every router platform has its own scaling capacity. The peak utilization will differ platform to platform.

How to check route flaps/network changes

Collect the following output to check the frequent route changes:

show route summary

start shell pfe network cfeb0

show heap 0
show heap 0 accounting pc size
show heap 1
show heap 1 accounting pc size
show piles
show nhdb zones
show nhdb management all
show jtree 0 memory

The next-hop statistics output of the command show nhdb management all will show the changes and deletion of routes. If  the counters increase frequently, then there are frequent route flaps. Users must then identify the source of the flaps.

Next-Hop Statistics:

Type         Adds       Changes    Deletes    Failures
------------ ---------- ---------- ---------- ----------
Discard      733        0          479        0
Reject       281        0          23         0
Unicast      1322468    5398       1317253    0
Unilist      719957     0          717624     0
Indexed      0          0          0          0
Indirect     165445     1418660    162924     0
Hold         2764506    378        2764502    0