Monitor Overall CPU utilisation

[EtherNation_Use]

Create Date: Jan 16 2013 10:31AM

We have a number of Extreme 450 stacks that are over time increasing traffic throughput on, with some 10G modules recently added which will increase overall traffic through the switch significantly. As part of monitoring the switch stack we want to understand how overall CPU utilisation is trending over time and planned to take some snapshots using show cpu-monitoring. Does the output of show cpu-monitoring show overall CPU utilisation on the first System lines shown for each slot or is that a specific process within the overall list ? If it's a summary total that's easier than exporting the output into an excel and totalling all the columns. The CLI reference guide doesn't provide any guidance either way on this, hence the question here.

As a separate task we are looking at SNMP to provide this but that also seems to have its challenges on overall CPU reporting.

Thanks for any guidance,

Euan

Slot-1 # show cpu

CPU Utilization Statistics - Monitored every 20 seconds
-------------------------------------------------------------------------------

Card Process 5 10 30 1 5 30 1 Max Total
secs secs secs min mins mins hour User/System
util util util util util util util util CPU Usage
(%) (%) (%) (%) (%) (%) (%) (%) (secs)
-------------------------------------------------------------------------------

Slot-1 System n/a n/a 3.6 2.7 2.9 3.0 2.9 75.8 3230.80 1494743.53
Slot-2 System n/a n/a 4.4 6.0 5.8 5.8 5.7 83.7 0.00 0.00

Slot-1 aaa n/a n/a 0.0 0.1 0.0 0.0 0.0 1.8 535.77 694.59
Slot-1 acl n/a n/a 0.0 0.2 0.1 0.1 0.1 1.4 23325.47 25110.31
Slot-1 bfd n/a n/a 0.0 0.4 0.3 0.3 0.3 4.2 18179.63 85851.05
Slot-1 bgp n/a n/a 0.0 0.0 0.0 0.0 0.0 1.3 4.08 2.65
Slot-1 brm n/a n/a 0.0 0.0 0.0 0.0 0.0 0.7 2.85 1.67
Slot-1 ces n/a n/a 0.0 0.0 0.0 0.0 0.0 0.0 0.00 0.00
Slot-1 cfgmgr n/a n/a 0.0 0.0 0.0 0.0 0.0 22.3 848.54 366.73
Slot-1 cli n/a n/a 0.0 0.0 0.0 1.1 0.5 93.9 7140.78 779.12
Slot-1 devmgr n/a n/a 0.0 0.0 0.0 0.0 0.0 6.6 1889.37 685.02
Slot-1 dirser n/a n/a 0.0 0.0 0.0 0.0 0.0 17.2 6.34 17.76
Slot-1 dosprotect n/a n/a 0.0 0.0 0.0 0.0 0.0 0.5 1.42 1.23
Slot-1 dot1ag n/a n/a 0.0 0.0 0.0 0.0 0.0 11.1 119.20 166.29
Slot-1 eaps n/a n/a 0.0 0.0 0.0 0.0 0.0 0.4 9.69 8.19
Slot-1 edp n/a n/a 0.0 0.0 0.0 0.0 0.0 6.7 747.86 440.36
............. (from euanb)

[EtherNation_Use]

Create Date: Jan 17 2013 7:49PM

Thanks for the information, that background is useful to know.

Does the System output per switch on the first lines of the show cpu-monitoring command equal the total across all processes or is it just one of the processes ?

regards

Euan (from euanb)

[EtherNation_Use]

Create Date: Jan 16 2013 1:45PM

Hello Euanb

the CPU of the switch is only used for communication to the switch (telnet snmp etc) as well as dynamic protocols (OSPF RIP EAPS) and broadcast or unknown packets.

The CPU does not (should not ) forward data traffic. I say should as there are instances where there is a major broadcast storm, loop in the network, or if the HW tables are full there the switch will try and forward traffic via the CPU.

The command you have is very useful, other commands to look at are show IPstats (this is all traffic that the CPU is forwarding) and show Process which shows each process and whether it is active or in a state where the process needs restarted.

Hope that helps
P

(from Paul_Russo)