Transition Networks
S4224 Web User Guide
33558 Rev. C
Page 555 of 669
Troubleshooting High CPU Load Conditions
After the S4224 completes the boot process, the switch CPU performs two distinct functions
simultaneously: it runs the various system processes required for a networked switch, and it sends /
receives packets to / from the S4224 hardware. CPU load increases when a system process requires
more time or when more network packets are sent and received. Under normal operating conditions, the
CPU is busy at least 5 percent of the time.
Since background S4224 processes on its switch timers execute multiple times per second, the S4224
never reports CPU utilization at 0%, even for very simple deployments. Normal data traffic packet
switching is done in the S4224 hardware without involving the CPU, so it is not affected by an overly busy
CPU.
The CPU becomes too busy when it receives too many packets from the S4224 hardware or when a
system process consumes too much CPU time. When either of these functions uses CPU resources to
the detriment of the other, the CPU becomes "too busy". For example, if the CPU is receiving numerous
packets because of a broadcast storm on the network, it becomes so busy processing all of the packets
that other system processes do not have access to CPU resources.
In many instances, high CPU load is normal and does not cause network problems. High CPU utilization
becomes a problem when the S4224 fails to perform as expected. CPU utilization spikes caused by a
known network event or activity are not problems (even an 85% spike may be acceptable, depending on
the cause).
Over time, the switch operates within a certain sustained CPU load range, which is considered the normal
operations baseline. You can use the output of the
system load
CLI command or the
Monitor
>
System
>
CPU
Load menu path.
The CPU Load percentage is shown at 100 ms, 1 second, and 10 second intervals. All numbers
represent running averages. Note that the web interface has an Auto-refresh checkbox to refresh the
page automatically every 3 seconds.
Frequent unexplained spikes to the established normal operating baseline, or sudden utilization jumps
with no explanations are likely causes for concern.
Below are some common symptoms of high CPU utilization.
High percentages in the CLI command output or web interface graph: check the output of the
system load
CLI command.
Slow performance: services fail to respond (e.g., slow Telnet response or unable to Telnet to the
S4224; slow console response, slow or no ping response).
If you notice any of these symptoms, follow the steps below to alleviate the problem.
1. Check for a possible security issue. A high CPU utilization can be caused by a security issue, such as
a worm or virus in your network. This is especially likely if there have not been recent changes to the
network. A configuration change (e.g., adding additional lines to ACLs) can mitigate the effects of this
problem.
2. Collect more information using the show version of the CLI commands (e.g., system config, system
log, etc.).
3. If the S4224 is accessible and you can reproduce the problem, try cycling power to the S4224.
4. Try lowering or disabling all sys logging. Increase logging buffer size.
5. Make sure any debug commands are turned off. Contact TN support for details.
