03-01-2012 02:07 PM - edited 03-10-2019 05:37 AM
We're getting ready to deploy 2 ASA 5585s with SSM-IPS40 modules. IDM reports 8 cpus at 100% utilization for a total of 50% cpu utilization. Is IDM reporting this in error? I'm not finding a way to check on the CLI. There is no traffic traversing these devices at the moment.
Thanks in advance for any help.
Dennis
Solved! Go to Solution.
03-05-2012 04:23 AM
As from E3 engine updates, the IPS uses a different algorithm for managing its idle
time,and spends more time polling for packets to reduce latency. The E3 changes included a
fix to a problem with high latency during low traffic loads. The fix was to have sensorApp
check the packet buffers on the driver more often. So the packet could be pulled off the
driver queue quicker for analysis instead of waiting for the driver to fill the queue
before passing it to sensorApp. This increased checking caused a corresponding increase in
cpu usage.
The CPU Utilization measurement can/will be high, due to the change made in
the E3 engine, polling the sensing interfaces for more data. This is OK as if/when the
sensorApp process requires additional CPU time, the amount the polling is using is
dynamically scaled back to useful for determining actual sensor load.
So do not worry for the HIGH CPU but check the Processing load percentage on the IPS.
Please post if there are still unanswered questions.
,
Sachin
03-05-2012 05:21 AM
show statistics virtual-sensor VS0
Virtual Sensor Statistics
Statistics for Virtual Sensor vs0
Processing Load Percentage = XX
or on the IDM where you see the Graph for the CPU in the same circlr you would see Inspection Load.
Sachin
03-05-2012 06:59 AM
Please mark as answered if answered
03-01-2012 02:38 PM
I should include the SSP_IPS40 blades shipped with 7.1(1)E4.
03-05-2012 04:23 AM
As from E3 engine updates, the IPS uses a different algorithm for managing its idle
time,and spends more time polling for packets to reduce latency. The E3 changes included a
fix to a problem with high latency during low traffic loads. The fix was to have sensorApp
check the packet buffers on the driver more often. So the packet could be pulled off the
driver queue quicker for analysis instead of waiting for the driver to fill the queue
before passing it to sensorApp. This increased checking caused a corresponding increase in
cpu usage.
The CPU Utilization measurement can/will be high, due to the change made in
the E3 engine, polling the sensing interfaces for more data. This is OK as if/when the
sensorApp process requires additional CPU time, the amount the polling is using is
dynamically scaled back to useful for determining actual sensor load.
So do not worry for the HIGH CPU but check the Processing load percentage on the IPS.
Please post if there are still unanswered questions.
,
Sachin
03-05-2012 05:15 AM
How would I check processing load?
Dennis
03-05-2012 05:21 AM
show statistics virtual-sensor VS0
Virtual Sensor Statistics
Statistics for Virtual Sensor vs0
Processing Load Percentage = XX
or on the IDM where you see the Graph for the CPU in the same circlr you would see Inspection Load.
Sachin
03-05-2012 05:58 AM
Ah, very good.
Thanks a bunch.
Dennis
03-05-2012 06:59 AM
Please mark as answered if answered
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide