RMON_EVENT_fpc_high_heap_utilization

RMON_EVENT_fpc_high_heap_utilization

The problem related to this syslog message is described in the following sections:

The RMON_EVENT_fpc_high_heap_utilization message contains information about a FPC that is experiencing high memory utilization.

An alarm is triggered when the heap utilization crosses a threshold. The message logged will look similar to this:

Heap represents a percentage measurement of the FPC’s memory, including memory for the FPC’s kernel and the space available for processing packets. If the percentage is high, it may simply because you are using that much memory. It may also indicate a memory leak, which will continue to use memory until it is exhausted, causing a crash. Because memory leaks can have widely varying causes, it will require further investigation.

The output from show chassis fpc will show the current heap utilization. It is useful to gather this data over time in order to track issues of increasing memory usage. As a general rule, any heap utilization of 80% or less should not be cause for concern. Periodic spikes of over 80% that then return to less are also to be expected. Constant heap utilization of 80% or more is grounds for suspicion of a memory leak. To further investigate, examine the following outputs:

Note: Replace fpc0 with the FPC that is experiencing the high heap utilization.

These outputs should contain clues that will help determine if there is any hardware issue or software error occurring.

During a maintenance window, perform the following test:

  1. Get the baseline memory usage via the command: show chassis fpc
  2. Restart the FPC.
  3. Monitor the memory usage by running show chassis fpc. If the memory utilization is legitimate, you should see heap utilization return to high levels relatively soon, with the caveat that the traffic using the FPC will have been disconnected, and it may take some time for full utilization to return. If the utilization stays low, but slowly and consistently continues to rise, then we would suspect a memory leak.

About the author

Prasanna

Leave a Comment