CPU spike in user processes
-
I may not have this issue anymore as it went away after rebooting, but I was wondering if anybody has heard of this issue or maybe knows what might cause it.
I got a new pfsense box today and did a fresh install. I noticed that the monitoring graphs were completely at 0 for half the day. After running for about 5 hours and having a total cpu load of 0-1% on average, the monitoring graphs start working and the user cpu usage spikes up to ~20% and stays like that for about 3 hours. I sshed in and checked top, mbufs, states, interrupts, and couldn't see anything wrong (I'm not very familiar with systat output, but top only showed 1 process over 1% - php fpm). The system process cpu usage was still ~1%, but user cpu usage in top and other utilities all showed around 20%.
What is usually run by user processes, and does anybody know of any common goofs that could cause this?
Any insight would be greatly appreciated, thanks!
(P.S. My hardware configuration is Modem -> Protectli micro firewall (pf sense) -> asus RT-N66u (ddwrt) in ap mode)
(P.P.S I attached a screenshot of the processor graph - it looks like before rebooting, the # of processes kept slowly increasing. After rebooting, it's been at 126 processes for 9 hours)
 -
Difficult to say without more detail, but on smaller hardware, just watching the dashboard on its own will cause a spike in CPU usage because it takes a fair amount of CPU time to process all of the data required to draw the dashboard.
-
Difficult to say without more detail, but on smaller hardware, just watching the dashboard on its own will cause a spike in CPU usage because it takes a fair amount of CPU time to process all of the data required to draw the dashboard.
In other words, the act of measuring can change the results.
Even on my i5 quad-core, viewing the web front-end bumps the CPU from 300mhz to 800mhz-1600mhz due to increased CPU load.
