Processor utilization grows over months of use?



  • I just looked at the processor utilization of my pfsense box and was shocked to see the "nice" utilization had grown to be about 40-50% of my CPU.  I have a Via-based pfSense box that does modest OpenVPN duty.  The RRD graph shows it went from about 1-3% utilization to the 50% level over a period of months.  I rebooted the box and it fell back to the 1-3% level.

    Any ideas what's going on?




  • What would have been interesting was to see what process was utilizing the CPU, via top or similar.

    Could have been a hung process of some sorts.



  • @eirikz:

    What would have been interesting was to see what process was utilizing the CPU, via top or similar.

    Actually, I looked for that.  When I logged in via SSH, tho, I did not see the utilization shown on the graph.  I did see an instance of sh which was "niced" and it was showing something like 10% utilization, but that was probably my SSH session.

    I should mention that at no time did I suspect anything was wrong because my pfSense box was routing with no perceived degradation.  I only became nervous when I saw the RRD graph and rebooted the box.  It's now averaging 0.13% nice, 1% system, 0.25% user, and 0.62% interrupt.


  • Rebel Alliance Developer Netgate

    From the console, try:

    top -SH
    

    Instead of just top. That will show system threads and kernel threads.



  • @jimp:

    From the console, try:

    top -SH
    

    Instead of just top. That will show system threads and kernel threads.

    Good to know.  I'll try that in a few months if it happens again.



  • i get the same thing with my ALIX




  • I'm seeing this again on my Nokia ip330 running 1.2.3 nanobsd with no noticeable performance degradation.

    top -SH output:

    last pid:  3198;  load averages:  0.97,  0.56,  0.42   up 17+14:10:30  11:08:10
    82 processes:  3 running, 63 sleeping, 2 zombie, 14 waiting
    CPU:  0.4% user, 32.7% nice, 21.5% system,  1.8% interrupt, 43.5% idle
    Mem: 40M Active, 19M Inact, 32M Wired, 72K Cache, 32M Buf, 147M Free
    Swap:
    
      PID USERNAME PRI NICE   SIZE    RES STATE    TIME   WCPU COMMAND
       10 root     171 ki31     0K     8K RUN    381.1H 61.96% idle
     9477 root       8   20  6564K  4896K wait   175:35  3.96% sh
       31 root     -68    -     0K     8K WAIT   121:58  0.00% irq5: fxp2
       12 root     -32    -     0K     8K RUN     63:02  0.00% swi4: clock sio
       29 root     -68    -     0K     8K WAIT    48:17  0.00% irq10: fxp0
     9615 nobody    64   20  3156K  1232K select  44:17  0.00% apinger
       14 root     -16    -     0K     8K -       18:31  0.00% yarrow
      944 nobody    44    0  3156K  1336K select  11:55  0.00% dnsmasq
       30 root     -68    -     0K     8K WAIT    11:22  0.00% irq12: fxp1
     1362 dhcpd     44    0  3156K  2164K select  10:51  0.00% dhcpd
       56 root      -8    -     0K     8K mdwait  10:29  0.00% md1
    14573 root      44    0  3524K  1868K RUN      5:12  0.00% top
      828 root       4    0  5144K  2952K kqread   3:35  0.00% lighttpd
        5 root      -8    -     0K     8K -        3:24  0.00% g_down
        3 root      -8    -     0K     8K -        3:22  0.00% g_event
        8 root       8    -     0K     8K -        3:19  0.00% thread taskq
    



Locked