05-05-2003 08:26 AM - edited 02-20-2020 10:43 PM
I found the following entry in our PIX log:
"...May 04 2003 02:24:49: %PIX-3-211003: CPU utilization for 10 seconds = 48366748%"
According to Cisco docs, this message displays if the CPU utilization is over 100% for 'x' number of seconds and the number after the equal sign is
"%cpu_usage".
So, does that mean our CPU went over 48 MILLION percent??? For 10 seconds???
If that's correct, what could cause this?
There were no unusual events at that time, according to the log. No unusual traffic amount.
Tony
05-05-2003 03:56 PM
This is a known issue and is purely cosmetic. See bug http://www.cisco.com/cgi-bin/Support/Bugtool/onebug.pl?bugid=CSCdz66410 for details. Basically if your PIX is lightly loaded for 49 straight days a counter wraps around and causes the error.
FYI, from the internal notes in the bug (you can't see these), the cause and fix is described as follows:
-----------------------------------------------------------
The elapsedtime field of Thread structure is a ulong.
If the system is lightly loaded, most of the time will be
spent in poll thread and the poll thread elapsedtime
field will wrap around. When this wrap happens, the
previous time will be more than the current time and
the CPU usage calculation results in a very high
usage.
On no load, it will wrap in 49 days.
Fixed the problem by changing the elapsedtime field from
ulong to uvlong. This will ensure that the wrap will not
happen in 584,942,417 years
------------------------------------------------------------------------
05-06-2003 04:55 AM
Thanks for your reply. I'll sleep better now.
Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: